One read of Claude Developer Platformapi-20261009T213714Z
59 pages moved out of 762 read.
Pages moved
59
significant first
Pages read
762
in this capture
Captured
21:37 UTC
Corpus hash
e3252cf6bcbe
index-hash
What this read moved
1-25 of 59, page 1 of 3This capture is too large to show at once. Changes 1-25 of 59 are below, significant first; the rest are on the following screens.
api/beta Changed · +5834 / -2504 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
from line 1544
15441544
15451545 Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages.
15461546
1547 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response.
1547 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. This is called prefill. On models that don't support prefill, creating a message that ends with a partial `assistant` response returns a 400 error. See [Prefill not supported](https://platform.claude.com/docs/en/api/errors#prefill-not-supported).
15481548
15491549 Example with a single `user` message:
15501550
from line 1562
15621562 ]
15631563 ```
15641564
1565 Example with a partially-filled response from Claude:
1565 Example with a partially-filled response from Claude, for models that support prefill:
15661566
15671567 ```json
15681568 [
from line 4549
45494549
45504550 - `"claude-haiku-4-5"`
45514551
4552 Fastest model with near-frontier intelligence
4553
45544552 - `"claude-haiku-4-5-20251001"`
45554553
4556 Fastest model with near-frontier intelligence
4557
45584554 - `"claude-opus-4-5"`
45594555
45604556 Powerful intelligence for long-running agents and coding
from line 5145
51495145
51505146- `fallbacks: optional BetaFallbacksParam or null`
51515147
5152 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. The string "default" requests the requested model's server-defined default fallback configuration.
5148 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. Some models don't support fallbacks; on those models, a list of fallback models returns a 400 error. See [Server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#server-side-fallback). The string "default" requests the requested model's server-defined default fallback configuration. On a model that doesn't support fallbacks, the request runs on the requested model alone, so a declined request stays declined.
51535149
51545150 - `array of BetaFallbackParam`
51555151
from line 5243
52475243
52485244 - `display: optional "summarized" or "omitted" or "updates" or null`
52495245
5250 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
5246 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
52515247
52525248 - `"summarized"`
52535249
from line 5269
52735269
52745270 - `display: optional "summarized" or "omitted" or "updates" or null`
52755271
5276 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
5272 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
52775273
52785274 - `"summarized"`
52795275
from line 5375
53795375
53805376- `thinking: optional BetaThinkingConfigParam`
53815377
5382 Configuration for enabling Claude's extended thinking.
5378 Configuration for Claude's thinking.
53835379
5384 When enabled, responses include `thinking` content blocks showing Claude's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit.
5380 With `{"type": "adaptive"}`, Claude decides when and how much to think. With `{"type": "enabled"}` (manual extended thinking), you set a `budget_tokens` of at least 1,024. Thinking tokens count toward your `max_tokens` limit.
53855381
5386 See [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for details.
5382 Which `type` values are accepted, and what happens when you omit `thinking`, depend on the model. See [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking#configuring-thinking) for each model's behavior.
53875383
53885384 - `BetaThinkingConfigEnabled object`
53895385
from line 7725
77297725
77307726 Name of the tool.
77317727
7732 This is how the tool will be called by the model and in `tool_use` blocks.
7733
7734 - `allowed_callers: optional array of "direct" or "code_execution_20250825" or "code_execution_20260120" or "code_execution_20260521"`
7735
7736 - `"direct"`
7737
7738 - `"code_execution_20250825"`
7739
7740 - `"code_execution_20260120"`
7741
7742 - `"code_execution_20260521"`
7743
7744 - `cache_control: optional BetaCacheControlEphemeral or null`
7745
7746 Create a cache control breakpoint at this content block.
7747
7748 - `defer_loading: optional boolean`
7749
7750 If true, tool will not be included in initial system prompt. Only loaded when returned via tool_reference from tool search.
7751
7752 - `input_examples: optional array of map[unknown]`
7753
7754 - `strict: optional boolean`
7755
7756 When true,
7728 This is how the tool will be called by
api/beta/agents Changed · +3367 / -1489 lines
### Beta Managed Agents Multiagent Advisor ### Beta Managed Agents Multiagent Advisor Disabled ### Beta Managed Agents Multiagent Advisor Disabled Params ### Beta Managed Agents Multiagent Advisor Enabled ### Beta Managed Agents Multiagent Advisor Enabled Params ### Beta Managed Agents Multiagent Advisor Params ### Beta Managed Agents Multiagent Inline Agents ### Beta Managed Agents Multiagent Inline Agents Disabled ### Beta Managed Agents Multiagent Inline Agents Disabled Params ### Beta Managed Agents Multiagent Inline Agents Enabled ### Beta Managed Agents Multiagent Inline Agents Enabled Params ### Beta Managed Agents Multiagent Inline Agents Params ### Beta Managed Agents Multiagent Predefined Agent Params ### Beta Managed Agents Multiagent Subagents ### Beta Managed Agents Multiagent Subagents Disabled ### Beta Managed Agents Multiagent Subagents Disabled Params ### Beta Managed Agents Multiagent Subagents Enabled ### Beta Managed Agents Multiagent Subagents Enabled Params ### Beta Managed Agents Multiagent Subagents Params ### Beta Managed Agents Multiagent Workflows ### Beta Managed Agents Multiagent Workflows Disabled ### Beta Managed Agents Multiagent Workflows Disabled Params ### Beta Managed Agents Multiagent Workflows Enabled ### Beta Managed Agents Multiagent Workflows Enabled Params ### Beta Managed Agents Multiagent Workflows Params ### Beta Managed Agents Multiagent20261001 ### Beta Managed Agents Multiagent20261001 Params
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 185
185185
186186 - `"claude-haiku-4-5"`
187187
188 Fastest model with near-frontier intelligence
189
190188 - `"claude-haiku-4-5-20251001"`
191189
192 Fastest model with near-frontier intelligence
193
194190 - `"claude-opus-4-5"`
195191
196192 Powerful intelligence for long-running agents and coding
from line 323
327323
328324 Multiagent orchestration configuration.
329325
330 - `type: "coordinator"`
331
332 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
333
334 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
335
336 - `string`
337
338 - `BetaManagedAgentsAgentParams object`
339
340 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
341
342 - `type: "agent"`
343
344 - `id: string`
345
346 The `agent` ID.
326 - `BetaManagedAgentsMultiagentCoordinatorParams object`
327
328 A coordinator topology: the session's primary thread orchestrates work by spawning session threads, each running an agent drawn from the `agents` roster.
329
330 - `type: "coordinator"`
331
332 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
333
334 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
335
336 - `string`
337
338 - `BetaManagedAgentsAgentParams object`
339
340 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
341
342 - `type: "agent"`
343
344 - `id: string`
345
346 The `agent` ID.
347
348 minLength: 1, maxLength: 128
349
350 - `version: optional number`
351
352 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
353
354 format: int32
355
356 - `BetaManagedAgentsMultiagentSelfParams object`
357
358 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
359
360 - `type: "self"`
361
362 - `BetaManagedAgentsAdvisorParams object`
363
364 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
365
366 - `type: "advisor"`
367
368 - `model: string`
369
370 A Claude model id. The model must be permitted as an advisor for this agent's model.
371
372 minLength: 1, maxLength: 256
373
374 - `BetaManagedAgentsMultiagent20261001Params object`
375
376 Multiagent configuration with three members, each enabled or disabled on its own. On an update, if the agent's stored `multiagent` also has type `multiagent_20261001`, this configuration is merged into the stored one, level by level, instead of replacing it. A key that the update omits keeps its stored value. A key sent as null takes its default, on create as well, so `"workflows": null` enables workflows. An object sent with a `type` other than the stored one replaces the stored object, and the keys that it omits take their defaults. A `predefined_agents` list that is sent replaces the stored list. Every object that is sent needs its `type`, and an enabled `advisor` needs its `model`. Other validation applies to the merged result.
377
378 - `type: "multiagent_20261001"`
379
380 - `advisor: optional BetaManagedAgentsMultiagentAdvisorParams or null`
381
382 Whether the session's primary thread can consult an advisor model. Defaults to disabled.
383
384 - `BetaManagedAgentsMultiagentAdvisorEnabledParams object`
385
386 The session's primary thread can consult `model` mid-turn.
387
388 - `type: "enabled"`
389
390 - `model: string`
391
392 A Claude model id. The model must be permitted as an advisor for this agent's model.
393
394 minLength: 1, maxLength: 256
395
396 - `BetaManagedAgentsMultiagentAdvisorDisabledParams object`
397
398 The agent has no advisor.
399
400 - `type: "disabled"`
401
402 - `subagents: optional BetaManagedAgentsMultiagentSubagentsParams or null`
403
404 Whether the agent can spawn session threads. Defaults to enabled.
405
406 - `BetaManagedAgentsMultiagentSubagentsEnabledParams object`
407
408 The agent can spawn session threads. Each thread runs a predefined agent, which is a saved agent in `predefined_agents`, or an inline agent, which the agent defines when it spawns the thread and which is not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
409
410 - `type: "enabled"`
411
412 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
413
414 Whether the agent can define inline agents when it spawns session threads. Defaults to enabled.
415
416 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
417
418 The agent can define inline agents.
419
420 - `type: "enabled"`
421
422 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
423
424 The agent cannot define inline agents.
425
426 - `type: "disabled"`
427
428 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
429
430 Predefined agents that this agent can spawn as session threads. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `workflows.predefined_agents`, and an agent in one list is not added to the other.
431
432 - `string`
433
434 - `BetaManagedAgentsAgentParams object`
435
436 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
437
438 - `BetaManagedAgentsMultiagentSelfParams object`
439
440 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
441
442 - `BetaManagedAgentsMultiagentSubagentsDisabledParams object`
443
444 The agent cannot spawn session threads.
445
446 - `type: "disabled"`
447
448 - `workflows: optional BetaManagedAgentsMultiagentWorkflowsParams or null`
449
450 Whether the agent can start workflow runs. Defaults to enabled.
451
452 - `BetaManagedAgentsMultiagentWorkflowsEnabledParams object`
453
454 The agent can start workflow runs. Each run follows a plan, a program that the agent writes. A plan can use predefined agents, which are the saved agents in `predefined_agents`, and inline agents, which it defines itself and which are not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
455
456 - `type: "enabled"`
457
458 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
459
460 Whether a run's plan can define inline agents. Defaults to enabled.
461
462 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
463
464 Predefined agents that a run's plan can use. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `subagents.predefined_agents`, and an agent in one list is not added to the other.
465
466 - `string`
467
468 - `BetaManagedAgentsAgentParams object`
469
470 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
471
472 - `BetaManagedAgentsMultiagentSelfParams object`
473
474 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
475
476 - `BetaManagedAgentsMultiagentWorkflowsDisabledParams object`
477
478 The agent cannot start workflow runs.
479
480 - `type: "disabled"`
481
482- `skills: optional array of BetaManagedAgentsSkillParams`
483
484 Skills available to the agent.
485
486 - `BetaManagedAgentsAnthropicSkillParams object`
487
488 An Anthropic-managed skill.
489
490 - `type: "anthropic"`
491
492 - `skill_id: string`
493
494 Identifier of the Anthropic skill (e.g., "xlsx").
495
496 minLength: 1, maxLength: 64
497
498 - `version: optional string or null`
499
500 Version to pin. Defaults to latest if omitted.
501
502 minLength: 1, maxLength: 64
503
504 - `BetaManagedAgentsCustomSkillParams object`
505
506 A user-created custom skill.
507
508 - `type: "custom"`
509
510 - `skill_id: string`
511
512 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
513
514 minLength: 1, maxLength: 64
515
516 - `version: optional string or null`
517
518 Version to pin. Defaults to latest if omitted.
519
520 minLength: 1, maxLength: 64
521
522- `system: optional string or null`
523
524 System prompt for the agent.
525
526 maxLength: 100000
527
528- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams`
529
530 Tool configurations available to the agent. Maximum of 256 tools across all toolsets allowed.
531
532 - `BetaManagedAgentsAgentToolset20260401Params object`
533
534 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
535
536 - `type: "agent_toolset_20260401"`
537
538 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
539
540 Per-tool configuration overrides.
541
542 - `BetaManagedAgentsBashToolConfigParams object`
543
544 Configuration override for the bash tool.
545
546 - `type: optional "bash"`
547
548 - `name: "bash"`
549
550 Must be "bash".
551
552 - `enabled: optional boolean or null`
553
554 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
555
556 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
557
558 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
559
560 - `BetaManagedAgentsAlwaysAllowPolicy object`
561
562 Tool calls are automatically approved without user confirmation.
563
564 - `type: "always_allow"`
565
566 - `BetaManagedAgentsAlwaysAskPolicy object`
567
568 Tool calls require user confirmation before execution.
569
570 - `type: "always_ask"`
571
572 - `BetaManagedAgentsAutoPolicy object`
573
574 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
575
576 - `type: "auto"`
577
578 - `BetaManagedAgentsEditToolConfigParams object`
579
580 Configuration override for the edit tool.
581
582 - `type: optional "edit"`
583
584 - `name: "edit"`
585
586 Must be "edit".
587
588 - `enabled: optional boolean or null`
589
590 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
591
592 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
593
594 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
595
596 - `BetaManagedAgentsAlwaysAllowPolicy object`
597
598 Tool calls are automatically approved without user confirmation.
599
600 - `BetaManagedAgentsAlwaysAskPolicy object`
601
602 Tool calls require user confirmation before execution.
603
604 - `BetaManagedAgentsAutoPolicy object`
605
606 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
607
608 - `BetaManagedAgentsReadToolConfigParams object`
609
610 Configuration override for the read tool.
611
612 - `type: optional "read"`
613
614 - `name: "read"`
615
616 Must be "read".
617
618 - `enabled: optional boolean or null`
619
620 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
621
622 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
623
624 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
625
626 - `BetaManagedAgentsAlwaysAllowPolicy object`
627
628 Tool calls are automatically approved without user confirmation.
629
630 - `BetaManagedAgentsAlwaysAskPolicy object`
631
632 Tool calls require user confirmation before execution.
633
634 - `BetaManagedAgentsAutoPolicy object`
635
636 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
637
638 - `BetaManagedAgentsWriteToolConfigParams object`
639
640 Configuration override for the write tool.
641
642 - `type: optional "write"`
643
644 - `name: "write"`
645
646 Must be "write".
647
648 - `enabled: optional boolean or null`
649
650 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
651
652 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
653
654 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
655
656 - `BetaManagedAgentsAlwaysAllowPolicy object`
657
658 Tool calls are automatically approved without user confirmation.
659
660 - `BetaManagedAgentsAlwaysAskPolicy object`
661
662 Tool calls require user confirmation before execution.
663
664 - `BetaManagedAgentsAutoPolicy object`
665
666 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
667
668 - `BetaManagedAgentsGlobToolConfigParams object`
669
670 Configuration override for the glob tool.
671
672 - `type: optional "glob"`
673
674 - `name: "glob"`
675
676 Must be "glob".
677
678 - `enabled: optional boolean or null`
679
680 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
681
682 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
683
684 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
685
686 - `BetaManagedAgentsAlwaysAllowPolicy object`
687
688 Tool calls are automatically approved without user confirmation.
689
690 - `BetaManagedAgentsAlwaysAskPolicy object`
691
692 Tool calls require user confirmation before execution.
693
694 - `BetaManagedAgentsAutoPolicy object`
695
696 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
697
698 - `BetaManagedAgentsGrepToolConfigParams object`
699
700 Configuration override for the grep tool.
701
702 - `type: optional "grep"`
703
704 - `name: "grep"`
705
706 Must be "grep".
707
708 - `enabled: optional boolean or null`
709
710 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
711
712 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
713
714 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
715
716 - `BetaManagedAgentsAlwaysAllowPolicy object`
717
718 Tool calls are automatically approved without user confirmation.
719
720 - `BetaManagedAgentsAlwaysAskPolicy object`
721
722 Tool calls require user confirmation before execution.
723
724 - `BetaManagedAgentsAutoPolicy object`
725
726 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
727
728 - `BetaManagedAgentsWebFetchToolConfigParams object`
729
730 Configuration override for the web_fetch tool.
731
732 - `type: optional "web_fetch"`
733
734 - `name: "web_fetch"`
735
736 Must be "web_fetch".
737
738 - `allowed_domains: optional array of string`
739
740 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
741
742 - `blocked_domains: optional array of string`
743
744 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
745
746 - `enabled: optional boolean or null`
747
748 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
749
750 - `max_content_tokens: optional number or null`
751
752 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
753
754 format: int32
755
756 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
757
758 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
759
760 - `BetaManagedAgentsAlwaysAllowPolicy object`
761
762 Tool calls are automatically approved without user confirmation.
763
764 - `BetaManagedAgentsAlwaysAskPolicy object`
765
766 Tool calls require user confirmation before execution.
767
768 - `BetaManagedAgentsAutoPolicy object`
769
770 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
771
772 - `url_sources: optional BetaManagedAgentsWebFetchURLSourcesParams or null`
773
774 Which sources contribute URLs the tool may fetch. Omit to allow every source.
775
776 - `client_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
777
778 Which custom tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be a custom tool in the same tools array.
779
780 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
781
782 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
783
784 - `"all"`
785
786 - `"none"`
787
788 - `BetaManagedAgentsWebFetchURLSourceToolFilter = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone or BetaManagedAgentsWebFetchURLSourceOnly or BetaManagedAgentsWebFetchURLSourceExcept`
789
790 Which tools' results contribute URLs that may be fetched.
791
792 - `BetaManagedAgentsWebFetchURLSourceAll object`
793
794 Every URL from this source may be fetched. This is the default.
795
796 - `type: "all"`
797
798 - `BetaManagedAgentsWebFetchURLSourceNone object`
799
800 This source contributes no URLs that may be fetched.
801
802 - `type: "none"`
803
804 - `BetaManagedAgentsWebFetchURLSourceOnly object`
805
806 Only the named tools' results contribute URLs that may be fetched.
807
808 - `type: "only"`
809
810 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
811
812 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
813
814 - `type: "tool_reference"`
815
816 Must be "tool_reference".
817
818 - `name: string`
819
820 Name of the tool. Compared exactly, so upper and lower case letters are different.
821
822 minLength: 1, maxLength: 128
823
824 - `BetaManagedAgentsWebFetchURLSourceExcept object`
825
826 Every tool's results contribute URLs that may be fetched, except the named tools' results.
827
828 - `type: "except"`
829
830 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
831
832 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
833
834 - `type: "tool_reference"`
835
836 Must be "tool_reference".
837
838 - `name: string`
839
840 Name of the tool. Compared exactly, so upper and lower case letters are different.
841
842 minLength: 1, maxLength: 128
843
844 - `server_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
845
846 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be "web_search" or "web_fetch".
847
848 - `user_input: optional BetaManagedAgentsWebFetchURLSourceUserInputParams or null`
849
850 Whether URLs in the text of user messages may be fetched: "all" (the default) or "none".
851
852 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
853
854 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
855
856 - `BetaManagedAgentsWebFetchURLSourceUserInput = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone`
857
858 Whether URLs in the text of user messages may be fetched.
859
860 - `BetaManagedAgentsWebFetchURLSourceAll object`
861
862 Every URL from this source may be fetched. This is the default.
863
864 - `BetaManagedAgentsWebFetchURLSourceNone object`
865
866 This source contributes no URLs that may be fetched.
867
868 - `BetaManagedAgentsWebSearchToolConfigParams object`
869
870 Configuration override for the web_search tool.
871
872 - `type: optional "web_search"`
873
874 - `name: "web_search"`
875
876 Must be "web_search".
877
878 - `allowed_domains: optional array of string`
879
880 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
881
882 - `blocked_domains: optional array of string`
883
884 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
885
886 - `enabled: optional boolean or null`
887
888 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
889
890 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
891
892 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
893
894 - `BetaManagedAgentsAlwaysAllowPolicy object`
895
896 Tool calls are automatically approved without user confirmation.
897
898 - `BetaManagedAgentsAlwaysAskPolicy object`
899
900 Tool calls require user confirmation before execution.
901
902 - `BetaManagedAgentsAutoPolicy object`
903
904 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
905
906 - `user_location: optional BetaManagedAgentsUserLocation or null`
907
908 Approximate user location for search result localization.
909
910 - `type: "approximate"`
911
912 Location precision. Only "approximate" is supported.
913
914 - `city: optional string or null`
915
916 City name.
917
918 minLength: 1, maxLength: 255
919
920 - `country: optional string or null`
921
922 Two-letter ISO 3166-1 country code, uppercase.
923
924 - `region: optional string or null`
925
926 Region or state name.
927
928 minLength: 1, maxLength: 255
929
930 - `timezone: optional string or null`
931
932 IANA timezone identifier, e.g. "America/Los_Angeles".
933
934 minLength: 1, maxLength: 255
935
936 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
937
938 Default configuration applied to all tools in this set.
939
940 - `enabled: optional boolean or null`
941
942 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
943
944 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
945
946 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
947
948 - `BetaManagedAgentsAlwaysAllowPolicy object`
949
950 Tool calls are automatically approved without user confirmation.
951
952 - `BetaManagedAgentsAlwaysAskPolicy object`
953
954 Tool calls require user confirmation before execution.
955
956 - `BetaManagedAgentsAutoPolicy object`
957
958 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
959
960 - `BetaManagedAgentsMCPToolsetParams object`
961
962 Configuration for tools from an MCP server defined in `mcp_servers`.
963
964 - `type: "mcp_toolset"`
965
966 - `mcp_server_name: string`
967
968 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
969
970 minLength: 1, maxLength: 255
971
972 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
973
974 Per-tool configuration overrides.
975
976 - `name: string`
977
978 Name of the MCP tool to configure. 1-128 characters.
347979
348980 minLength: 1, maxLength: 128
349981
350 - `version: optional number`
351
352 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
353
354 format: int32
355
356 - `BetaManagedAgentsMultiagentSelfParams object`
357
358 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
359
360 - `type: "self"`
361
362 - `BetaManagedAgentsAdvisorParams object`
363
364 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
365
366 - `type: "advisor"`
367
368 - `model: string`
369
370 A Claude model id. The model must be permitted as an advisor for this agent's model.
371
372 minLength: 1, maxLength: 256
373
374- `skills: optional array of BetaManagedAgentsSkillParams`
375
376 Skills available to the agent.
377
378 - `BetaManagedAgentsAnthropicSkillParams object`
379
380 An Anthropic-managed skill.
381
382 - `type: "anthropic"`
383
384 - `skill_id: string`
385
386 Identifier of the Anthropic skill (e.g., "xlsx").
387
388 minLength: 1, maxLength: 64
389
390 - `version: optional string or null`
391
392 Version to pin. Defaults to latest if omitted.
393
394 minLength: 1, maxLength: 64
395
396 - `BetaManagedAgentsCustomSkillParams object`
397
398 A user-created custom skill.
982 - `enabled: optional boolean or null`
983
984 Whether this tool is enabled. Overrides the `default_config` setting.
985
986 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
987
988 Permission policy for this tool. Overrides the `default_config` setting.
989
990 - `BetaManagedAgentsAlwaysAllowPolicy object`
991
992 Tool calls are automatically approved without user confirmation.
993
994 - `BetaManagedAgentsAlwaysAskPolicy object`
995
996 Tool calls require user confirmation before execution.
997
998 - `BetaManagedAgentsAutoPolicy object`
999
1000 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1001
1002 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
1003
1004 Default configuration for all tools from this server.
1005
1006 - `enabled: optional boolean or null`
1007
1008 Whether tools are enabled by default. Defaults to true if not specified.
1009
1010 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
1011
1012 Default permission policy for tools from this server.
1013
1014 - `BetaManagedAgentsAlwaysAllowPolicy object`
1015
1016 Tool calls are automatically approved without user confirmation.
1017
1018 - `BetaManagedAgentsAlwaysAskPolicy object`
1019
1020 Tool calls require user confirmation before execution.
1021
1022 - `BetaManagedAgentsAutoPolicy object`
1023
1024 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1025
1026 - `BetaManagedAgentsCustomToolParams object`
1027
1028 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
3991029
4001030 - `type: "custom"`
4011031
402 - `skill_id: string`
403
404 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
405
406 minLength: 1, maxLength: 64
407
408 - `version: optional string or null`
409
410 Version to pin. Defaults to latest if omitted.
411
412 minLength: 1, maxLength: 64
413
414- `system: optional string or null`
415
416 System prompt for the agent.
417
418 maxLength: 100000
419
420- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams`
421
422 Tool configurations available to the agent. Maximum of 256 tools across all toolsets allowed.
423
424 - `BetaManagedAgentsAgentToolset20260401Params object`
425
426 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
427
428 - `type: "agent_toolset_20260401"`
429
430 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
431
432 Per-tool configuration overrides.
433
434 - `BetaManagedAgentsBashToolConfigParams object`
435
436 Configuration override for the bash tool.
437
438 - `type: optional "bash"`
439
440 - `name: "bash"`
441
442 Must be "bash".
443
444 - `enabled: optional boolean or null`
445
446 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
447
448 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
449
450 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
451
452 - `BetaManagedAgentsAlwaysAllowPolicy object`
453
454 Tool calls are automatically approved without user confirmation.
455
456 - `type: "always_allow"`
457
458 - `BetaManagedAgentsAlwaysAskPolicy object`
459
460 Tool calls require user confirmation before execution.
461
462 - `type: "always_ask"`
463
464 - `BetaManagedAgentsAutoPolicy object`
465
466 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
467
468 - `type: "auto"`
469
470 - `BetaManagedAgentsEditToolConfigParams object`
471
472 Configuration override for the edit tool.
473
474 - `type: optional "edit"`
475
476 - `name: "edit"`
477
478 Must be "edit".
479
480 - `enabled: optional boolean or null`
481
482 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
483
484 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
485
486 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
487
488 - `BetaManagedAgentsAlwaysAllowPolicy object`
489
490 Tool calls are automatically approved without user confirmation.
491
492 - `BetaManagedAgentsAlwaysAskPolicy object`
493
494 Tool calls require user confirmation before execution.
495
496 - `BetaManagedAgentsAutoPolicy object`
497
498 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
499
500 - `BetaManagedAgentsReadToolConfigParams object`
501
502 Configuration override for the read tool.
503
504 - `type: optional "read"`
505
506 - `name: "read"`
507
508 Must be "read".
509
510 - `enabled: optional boolean or null`
511
512 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
513
514 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
515
516 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
517
518 - `BetaManagedAgentsAlwaysAllowPolicy object`
519
520 Tool calls are automatically approved without user confirmation.
521
522 - `BetaManagedAgentsAlwaysAskPolicy object`
523
524 Tool calls require user confirmation before execution.
525
526 - `BetaManagedAgentsAutoPolicy object`
527
528 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
529
530 - `BetaManagedAgentsWriteToolConfigParams object`
531
532 Configuration override for the write tool.
533
534 - `type: optional "write"`
535
536 - `name: "write"`
537
538 Must be "write".
539
540 - `enabled: optional boolean or null`
541
542 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
543
544 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
545
546 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
547
548 - `BetaManagedAgentsAlwaysAllowPolicy object`
549
550 Tool calls are automatically approved without user confirmation.
551
552 - `BetaManagedAgentsAlwaysAskPolicy object`
553
554 Tool calls require user confirmation before execution.
555
556 - `BetaManagedAgentsAutoPolicy object`
557
558 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
559
560 - `BetaManagedAgentsGlobToolConfigParams object`
561
562 Configuration override for the glob tool.
563
564 - `type: optional "glob"`
565
566 - `name: "glob"`
567
568 Must be "glob".
569
570 - `enabled: optional boolean or null`
571
572 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
573
574 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
575
576 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
577
578 - `BetaManagedAgentsAlwaysAllowPolicy object`
579
580 Tool calls are automatically approved without user confirmation.
581
582 - `BetaManagedAgentsAlwaysAskPolicy object`
583
584 Tool calls require user confirmation before execution.
585
586 - `BetaManagedAgentsAutoPolicy object`
587
588 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
589
590 - `BetaManagedAgentsGrepToolConfigParams object`
591
592 Configuration override for the grep tool.
593
594 - `type: optional "grep"`
595
596 - `name: "grep"`
597
598 Must be "grep".
599
600 - `enabled: optional boolean or null`
601
602 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
603
604 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
605
606 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
607
608 - `BetaManagedAgentsAlwaysAllowPolicy object`
609
610 Tool calls are automatically approved without user confirmation.
611
612 - `BetaManagedAgentsAlwaysAskPolicy object`
613
614 Tool calls require user confirmation before execution.
615
616 - `BetaManagedAgentsAutoPolicy object`
617
618 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
619
620 - `BetaManagedAgentsWebFetchToolConfigParams object`
621
622 Configuration override for the web_fetch tool.
623
624 - `type: optional "web_fetch"`
625
626 - `name: "web_fetch"`
627
628 Must be "web_fetch".
629
630 - `allowed_domains: optional array of string`
631
632 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
633
634 - `blocked_domains: optional array of string`
635
636 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
637
638 - `enabled: optional boolean or null`
639
640 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
641
642 - `max_content_tokens: optional number or null`
643
644 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
645
646 format: int32
647
648 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
649
650 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
651
652 - `BetaManagedAgentsAlwaysAllowPolicy object`
653
654 Tool calls are automatically approved without user confirmation.
655
656 - `BetaManagedAgentsAlwaysAskPolicy object`
657
658 Tool calls require user confirmation before execution.
659
660 - `BetaManagedAgentsAutoPolicy object`
661
662 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
663
664 - `url_sources: optional BetaManagedAgentsWebFetchURLSourcesParams or null`
665
666 Which sources contribute URLs the tool may fetch. Omit to allow every source.
667
668 - `client_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
669
670 Which custom tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be a custom tool in the same tools array.
671
672 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
673
674 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
675
676 - `"all"`
677
678 - `"none"`
679
680 - `BetaManagedAgentsWebFetchURLSourceToolFilter = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone or BetaManagedAgentsWebFetchURLSourceOnly or BetaManagedAgentsWebFetchURLSourceExcept`
681
682 Which tools' results contribute URLs that may be fetched.
683
684 - `BetaManagedAgentsWebFetchURLSourceAll object`
685
686 Every URL from this source may be fetched. This is the default.
687
688 - `type: "all"`
689
690 - `BetaManagedAgentsWebFetchURLSourceNone object`
691
692 This source contributes no URLs that may be fetched.
693
694 - `type: "none"`
695
696 - `BetaManagedAgentsWebFetchURLSourceOnly object`
697
698 Only the named tools' results contribute URLs that may be fetched.
699
700 - `type: "only"`
701
702 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
703
704 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
705
706 - `type: "tool_reference"`
707
708 Must be "tool_reference".
709
710 - `name: string`
711
712 Name of the tool. Compared exactly, so upper and lower case letters are different.
713
714 minLength: 1, maxLength: 128
715
716 - `BetaManagedAgentsWebFetchURLSourceExcept object`
717
718 Every tool's results contribute URLs that may be fetched, except the named tools' results.
719
720 - `type: "except"`
721
722 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
723
724 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
725
726 - `type: "tool_reference"`
727
728 Must be "tool_reference".
729
730 - `name: string`
731
732 Name of the tool. Compared exactly, so upper and lower case letters are different.
733
734 minLength: 1, maxLength: 128
735
736 - `server_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
737
738 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be "web_search" or "web_fetch".
739
740 - `user_input: optional BetaManagedAgentsWebFetchURLSourceUserInputParams or null`
741
742 Whether URLs in the text of user messages may be fetched: "all" (the default) or "none".
743
744 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
745
746 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
747
748 - `BetaManagedAgentsWebFetchURLSourceUserInput = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone`
749
750 Whether URLs in the text of user messages may be fetched.
751
752 - `BetaManagedAgentsWebFetchURLSourceAll object`
753
754 Every URL from this source may be fetched. This is the default.
755
756 - `BetaManagedAgentsWebFetchURLSourceNone object`
757
758 This source contributes no URLs that may be fetched.
759
760 - `BetaManagedAgentsWebSearchToolConfigParams object`
761
762 Configuration override for the web_search tool.
763
764 - `type: optional "web_search"`
765
766 - `name: "web_search"`
767
768 Must be "web_search".
769
770 - `allowed_domains: optional array of string`
771
772 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
773
774 - `blocked_domains: optional array of string`
775
776 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
777
778 - `enabled: optional boolean or null`
779
780 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
781
782 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
783
784 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
785
786 - `BetaManagedAgentsAlwaysAllowPolicy object`
787
788 Tool calls are automatically approved without user confirmation.
789
790 - `BetaManagedAgentsAlwaysAskPolicy object`
791
792 Tool calls require user confirmation before execution.
793
794 - `BetaManagedAgentsAutoPolicy object`
795
796 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
797
798 - `user_location: optional BetaManagedAgentsUserLocation or null`
799
800 Approximate user location for search result localization.
801
802 - `type: "approximate"`
803
804 Location precision. Only "approximate" is supported.
805
806 - `city: optional string or null`
807
808 City name.
809
810 minLength: 1, maxLength: 255
811
812 - `country: optional string or null`
813
814 Two-letter ISO 3166-1 country code, uppercase.
815
816 - `region: optional string or null`
817
818 Region or state name.
819
820 minLength: 1, maxLength: 255
821
822 - `timezone: optional string or null`
823
824 IANA timezone identifier, e.g. "America/Los_Angeles".
825
826 minLength: 1, maxLength: 255
827
828 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
829
830 Default configuration applied to all tools in this set.
831
832 - `enabled: optional boolean or null`
833
834 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
835
836 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
837
838 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
839
840 - `BetaManagedAgentsAlwaysAllowPolicy object`
841
842 Tool calls are automatically approved without user confirmation.
843
844 - `BetaManagedAgentsAlwaysAskPolicy object`
845
846 Tool calls require user confirmation before execution.
847
848 - `BetaManagedAgentsAutoPolicy object`
849
850 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
851
852 - `BetaManagedAgentsMCPToolsetParams object`
853
854 Configuration for tools from an MCP server defined in `mcp_servers`.
855
856 - `type: "mcp_toolset"`
857
858 - `mcp_server_name: string`
859
860 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
861
862 minLength: 1, maxLength: 255
863
864 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
865
866 Per-tool configuration overrides.
867
868 - `name: string`
869
870 Name of the MCP tool to configure. 1-128 characters.
871
872 minLength: 1, maxLength: 128
873
874 - `enabled: optional boolean or null`
875
876 Whether this tool is enabled. Overrides the `default_config` setting.
877
878 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
879
880 Permission policy for this tool. Overrides the `default_config` setting.
881
882 - `BetaManagedAgentsAlwaysAllowPolicy object`
883
884 Tool calls are automatically approved without user confirmation.
885
886 - `BetaManagedAgentsAlwaysAskPolicy object`
887
888 Tool calls require user confirmation before execution.
889
890 - `BetaManagedAgentsAutoPolicy object`
891
892 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
893
894 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
895
896 Default configuration for all tools from this server.
897
898 - `enabled: optional boolean or null`
899
900 Whether tools are enabled by default. Defaults to true if not specified.
901
902 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
903
904 Default permission policy for tools from this server.
905
906 - `BetaManagedAgentsAlwaysAllowPolicy object`
907
908 Tool calls are automatically approved without user confirmation.
909
910 - `BetaManagedAgentsAlwaysAskPolicy object`
911
912 Tool calls require user confirmation before execution.
913
914 - `BetaManagedAgentsAutoPolicy object`
915
916 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
917
918 - `BetaManagedAgentsCustomToolParams object`
919
920 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
921
922 - `type: "custom"`
923
9241032 - `description: string`
9251033
9261034 Description of what the tool does, shown to the agent to help it decide when to use the tool.
from line 1141
10331141
10341142 - `"claude-haiku-4-5"`
10351143
1036 Fastest model with near-frontier intelligence
1037
10381144 - `"claude-haiku-4-5-20251001"`
10391145
1040 Fastest model with near-frontier intelligence
1041
10421146 - `"claude-opus-4-5"`
10431147
10441148 Powerful intelligence for long-running agents and coding
from line 1215
11111215
11121216 Multiagent orchestration configuration. Null when the agent is single-threaded.
11131217
1114 - `type: "coordinator"`
1115
1116 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
1117
1118 Agents the coordinator may spawn as session threads, each resolved to a specific version.
1119
1120 - `BetaManagedAgentsAgentReference object`
1121
1122 A resolved agent reference with a concrete version.
1123
1124 - `type: "agent"`
1125
1126 - `id: string`
1127
1128 - `version: number`
1129
1130 format: int32
1131
1132 - `BetaManagedAgentsAdvisor object`
1133
1134 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
1135
1136 - `type: "advisor"`
1137
1138 - `model: string`
1139
1140 The advisor model id.
1218 - `BetaManagedAgentsMultiagentCoordinator object`
1219
1220 Resolved coordinator topology with a concrete agent roster.
1221
1222 - `type: "coordinator"`
1223
1224 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
1225
1226 Agents the coordinator may spawn as session threads, each resolved to a specific version.
1227
1228 - `BetaManagedAgentsAgentReference object`
1229
1230 A resolved agent reference with a concrete version.
1231
1232 - `type: "agent"`
1233
1234 - `id: string`
1235
1236 - `version: number`
1237
1238 format: int32
1239
1240 - `BetaManagedAgentsAdvisor object`
1241
1242 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
1243
1244 - `type: "advisor"`
1245
1246 - `model: string`
1247
1248 The advisor model id.
1249
1250 - `BetaManagedAgentsMultiagent20261001 object`
1251
1252 Resolved multiagent configuration with three members, each enabled or disabled on its own.
1253
1254 - `type: "multiagent_20261001"`
1255
1256 - `advisor: BetaManagedAgentsMultiagentAdvisor`
1257
1258 Whether the session's primary thread can consult an advisor model.
1259
1260 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
1261
1262 The session's primary thread can consult `model` mid-turn.
1263
1264 - `type: "enabled"`
1265
1266 - `model: string`
1267
1268 The advisor model id.
1269
1270 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
1271
1272 The agent has no advisor.
1273
1274 - `type: "disabled"`
1275
1276 - `subagents: BetaManagedAgentsMultiagentSubagents`
1277
1278 Whether the agent can spawn session threads.
1279
1280 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
1281
1282 The agent can spawn session threads.
1283
1284 - `type: "enabled"`
1285
1286 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
1287
1288 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
1289
1290 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
1291
1292 The agent can define inline agents.
1293
1294 - `type: "enabled"`
1295
1296 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
1297
1298 The agent cannot define inline agents.
1299
1300 - `type: "disabled"`
1301
1302 - `predefined_agents: array of BetaManagedAgentsAgentReference`
1303
1304 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
1305
1306 - `type: "agent"`
1307
1308 - `id: string`
1309
1310 - `version: number`
1311
1312 format: int32
1313
1314 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
1315
1316 The agent cannot spawn session threads.
1317
1318 - `type: "disabled"`
1319
1320 - `workflows: BetaManagedAgentsMultiagentWorkflows`
1321
1322 Whether the agent can start workflow runs.
1323
1324 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
1325
1326 The agent can start workflow runs.
1327
1328 - `type: "enabled"`
1329
1330 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
1331
1332 Whether a run's plan can define inline agents, which are not saved.
1333
1334 - `predefined_agents: array of BetaManagedAgentsAgentReference`
1335
1336 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
1337
1338 - `type: "agent"`
1339
1340 - `id: string`
1341
1342 - `version: number`
1343
1344 format: int32
1345
1346 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
1347
1348 The agent cannot start workflow runs.
1349
1350 - `type: "disabled"`
11411351
11421352 - `name: string`
11431353
from line 1833
16231833 "description": "A general-purpose starter agent.",
16241834 "metadata": {
16251835 "foo": "bar"
1626 },
1627 "multiagent": {
1628 "agents": [
1629 "agent_011CZkYqphY8vELVzwCUpqiQ",
1630 {
1631 "type": "self"
1632 }
1633 ],
1634 "type": "coordinator"
16351836 },
16361837 "system": "You are a general-purpose agent that can research, write code, run commands, and use connected tools to complete the user'\''s task end to end.",
16371838 "tools": [
from line 1870
16691870 "speed": "standard"
16701871 },
16711872 "multiagent": {
1672 "agents": [
1673 {
1674 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1675 "type": "agent",
1676 "version": 1
1677 }
1678 ],
1679 "type": "coordinator"
1873 "advisor": {
1874 "type": "disabled"
1875 },
1876 "subagents": {
1877 "inline_agents": {
1878 "type": "enabled"
1879 },
1880 "predefined_agents": [
1881 {
1882 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1883 "type": "agent",
1884 "version": 1
1885 }
1886 ],
1887 "type": "enabled"
1888 },
1889 "type": "multiagent_20261001",
1890 "workflows": {
1891 "inline_agents": {
1892 "type": "enabled"
1893 },
1894 "predefined_agents": [
1895 {
1896 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1897 "type": "agent",
1898 "version": 1
1899 }
1900 ],
1901 "type": "enabled"
1902 }
16801903 },
16811904 "name": "My First Agent",
16821905 "skills": [
from line 2182
19592182
19602183 - `"claude-haiku-4-5"`
19612184
1962 Fastest model with near-frontier intelligence
1963
19642185 - `"claude-haiku-4-5-20251001"`
19652186
1966 Fastest model with near-frontier intelligence
1967
19682187 - `"claude-opus-4-5"`
19692188
19702189 Powerful intelligence for long-running agents and coding
from line 2256
20372256
20382257 Multiagent orchestration configuration. Null when the agent is single-threaded.
20392258
2040 - `type: "coordinator"`
2041
2042 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
2043
2044 Agents the coordinator may spawn as session threads, each resolved to a specific version.
2045
2046 - `BetaManagedAgentsAgentReference object`
2047
2048 A resolved agent reference with a concrete version.
2049
2050 - `type: "agent"`
2051
2052 - `id: string`
2053
2054 - `version: number`
2055
2056 format: int32
2057
2058 - `BetaManagedAgentsAdvisor object`
2059
2060 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
2061
2062 - `type: "advisor"`
2063
2064 - `model: string`
2065
2066 The advisor model id.
2259 - `BetaManagedAgentsMultiagentCoordinator object`
2260
2261 Resolved coordinator topology with a concrete agent roster.
2262
2263 - `type: "coordinator"`
2264
2265 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
2266
2267 Agents the coordinator may spawn as session threads, each resolved to a specific version.
2268
2269 - `BetaManagedAgentsAgentReference object`
2270
2271 A resolved agent reference with a concrete version.
2272
2273 - `type: "agent"`
2274
2275 - `id: string`
2276
2277 - `version: number`
2278
2279 format: int32
2280
2281 - `BetaManagedAgentsAdvisor object`
2282
2283 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
2284
2285 - `type: "advisor"`
2286
2287 - `model: string`
2288
2289 The advisor model id.
2290
2291 - `BetaManagedAgentsMultiagent20261001 object`
2292
2293 Resolved multiagent configuration with three members, each enabled or disabled on its own.
2294
2295 - `type: "multiagent_20261001"`
2296
2297 - `advisor: BetaManagedAgentsMultiagentAdvisor`
2298
2299 Whether the session's primary thread can consult an advisor model.
2300
2301 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
2302
2303 The session's primary thread can consult `model` mid-turn.
2304
2305 - `type: "enabled"`
2306
2307 - `model: string`
2308
2309 The advisor model id.
2310
2311 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
2312
2313 The agent has no advisor.
2314
2315 - `type: "disabled"`
2316
2317 - `subagents: BetaManagedAgentsMultiagentSubagents`
2318
2319 Whether the agent can spawn session threads.
2320
2321 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
2322
2323 The agent can spawn session threads.
2324
2325 - `type: "enabled"`
2326
2327 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
2328
2329 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
2330
2331 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
2332
2333 The agent can define inline agents.
2334
2335 - `type: "enabled"`
2336
2337 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
2338
2339 The agent cannot define inline agents.
2340
2341 - `type: "disabled"`
2342
2343 - `predefined_agents: array of BetaManagedAgentsAgentReference`
2344
2345 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
2346
2347 - `type: "agent"`
2348
2349 - `id: string`
2350
2351 - `version: number`
2352
2353 format: int32
2354
2355 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
2356
2357 The agent cannot spawn session threads.
2358
2359 - `type: "disabled"`
2360
2361 - `workflows: BetaManagedAgentsMultiagentWorkflows`
2362
2363 Whether the agent can start workflow runs.
2364
2365 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
2366
2367 The agent can start workflow runs.
2368
2369 - `type: "enabled"`
2370
2371 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
2372
2373 Whether a run's plan can define inline agents, which are not saved.
2374
2375 - `predefined_agents: array of BetaManagedAgentsAgentReference`
2376
2377 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
2378
2379 - `type: "agent"`
2380
2381 - `id: string`
2382
2383 - `version: number`
2384
2385 format: int32
2386
2387 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
2388
2389 The agent cannot start workflow runs.
2390
2391 - `type: "disabled"`
20672392
20682393 - `name: string`
20692394
from line 2902
25772902 "speed": "standard"
25782903 },
25792904 "multiagent": {
2580 "agents": [
2581 {
2582 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
2583 "type": "agent",
2584 "version": 1
2585 }
2586 ],
2587 "type": "coordinator"
2905 "advisor": {
2906 "type": "disabled"
2907 },
2908 "subagents": {
2909 "inline_agents": {
2910 "type": "enabled"
2911 },
2912 "predefined_agents": [
2913 {
2914 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
2915 "type": "agent",
2916 "version": 1
2917 }
2918 ],
2919 "type": "enabled"
2920 },
2921 "type": "multiagent_20261001",
2922 "workflows": {
2923 "inline_agents": {
2924 "type": "enabled"
2925 },
2926 "predefined_agents": [
2927 {
2928 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
2929 "type": "agent",
2930 "version": 1
2931 }
2932 ],
2933 "type": "enabled"
2934 }
25882935 },
25892936 "name": "My First Agent",
25902937 "skills": [
from line 3203
28563203
28573204 - `"claude-haiku-4-5"`
28583205
2859 Fastest model with near-frontier intelligence
2860
28613206 - `"claude-haiku-4-5-20251001"`
28623207
2863 Fastest model with near-frontier intelligence
2864
28653208 - `"claude-opus-4-5"`
28663209
28673210 Powerful intelligence for long-running agents and coding
from line 3277
29343277
29353278 Multiagent orchestration configuration. Null when the agent is single-threaded.
29363279
2937 - `type: "coordinator"`
2938
2939 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
2940
2941 Agents the coordinator may spawn as session threads, each resolved to a specific version.
2942
2943 - `BetaManagedAgentsAgentReference object`
2944
2945 A resolved agent reference with a concrete version.
2946
2947 - `type: "agent"`
2948
2949 - `id: string`
2950
2951 - `version: number`
2952
2953 format: int32
2954
2955 - `BetaManagedAgentsAdvisor object`
2956
2957 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
2958
2959 - `type: "advisor"`
2960
2961 - `model: string`
2962
2963 The advisor model id.
3280 - `BetaManagedAgentsMultiagentCoordinator object`
3281
3282 Resolved coordinator topology with a concrete agent roster.
3283
3284 - `type: "coordinator"`
3285
3286 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
3287
3288 Agents the coordinator may spawn as session threads, each resolved to a specific version.
3289
3290 - `BetaManagedAgentsAgentReference object`
3291
3292 A resolved agent reference with a concrete version.
3293
3294 - `type: "agent"`
3295
3296 - `id: string`
3297
3298 - `version: number`
3299
3300 format: int32
3301
3302 - `BetaManagedAgentsAdvisor object`
3303
3304 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
3305
3306 - `type: "advisor"`
3307
3308 - `model: string`
3309
3310 The advisor model id.
3311
3312 - `BetaManagedAgentsMultiagent20261001 object`
3313
3314 Resolved multiagent configuration with three members, each enabled or disabled on its own.
3315
3316 - `type: "multiagent_20261001"`
3317
3318 - `advisor: BetaManagedAgentsMultiagentAdvisor`
3319
3320 Whether the session's primary thread can consult an advisor model.
3321
3322 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
3323
3324 The session's primary thread can consult `model` mid-turn.
3325
3326 - `type: "enabled"`
3327
3328 - `model: string`
3329
3330 The advisor model id.
3331
3332 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
3333
3334 The agent has no advisor.
3335
3336 - `type: "disabled"`
3337
3338 - `subagents: BetaManagedAgentsMultiagentSubagents`
3339
3340 Whether the agent can spawn session threads.
3341
3342 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
3343
3344 The agent can spawn session threads.
3345
3346 - `type: "enabled"`
3347
3348 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
3349
3350 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
3351
3352 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
3353
3354 The agent can define inline agents.
3355
3356 - `type: "enabled"`
3357
3358 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
3359
3360 The agent cannot define inline agents.
3361
3362 - `type: "disabled"`
3363
3364 - `predefined_agents: array of BetaManagedAgentsAgentReference`
3365
3366 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
3367
3368 - `type: "agent"`
3369
3370 - `id: string`
3371
3372 - `version: number`
3373
3374 format: int32
3375
3376 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
3377
3378 The agent cannot spawn session threads.
3379
3380 - `type: "disabled"`
3381
3382 - `workflows: BetaManagedAgentsMultiagentWorkflows`
3383
3384 Whether the agent can start workflow runs.
3385
3386 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
3387
3388 The agent can start workflow runs.
3389
3390 - `type: "enabled"`
3391
3392 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
3393
3394 Whether a run's plan can define inline agents, which are not saved.
3395
3396 - `predefined_agents: array of BetaManagedAgentsAgentReference`
3397
3398 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
3399
3400 - `type: "agent"`
3401
3402 - `id: string`
3403
3404 - `version: number`
3405
3406 format: int32
3407
3408 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
3409
3410 The agent cannot start workflow runs.
3411
3412 - `type: "disabled"`
29643413
29653414 - `name: string`
29663415
from line 3917
34683917 "speed": "standard"
34693918 },
34703919 "multiagent": {
3471 "agents": [
3472 {
3473 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
3474 "type": "agent",
3475 "version": 1
3476 }
3477 ],
3478 "type": "coordinator"
3920 "advisor": {
3921 "type": "disabled"
3922 },
3923 "subagents": {
3924 "inline_agents": {
3925 "type": "enabled"
3926 },
3927 "predefined_agents": [
3928 {
3929 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
3930 "type": "agent",
3931 "version": 1
3932 }
3933 ],
3934 "type": "enabled"
3935 },
3936 "type": "multiagent_20261001",
3937 "workflows": {
3938 "inline_agents": {
3939 "type": "enabled"
3940 },
3941 "predefined_agents": [
3942 {
3943 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
3944 "type": "agent",
3945 "version": 1
3946 }
3947 ],
3948 "type": "enabled"
3949 }
34793950 },
34803951 "name": "My First Agent",
34813952 "skills": [
from line 4203
37324203
37334204 - `"claude-haiku-4-5"`
37344205
3735 Fastest model with near-frontier intelligence
3736
37374206 - `"claude-haiku-4-5-20251001"`
37384207
3739 Fastest model with near-frontier intelligence
3740
37414208 - `"claude-opus-4-5"`
37424209
37434210 Powerful intelligence for long-running agents and coding
from line 4307
38404307
38414308 Multiagent orchestration configuration. Full replacement. Omit to preserve; send null to clear.
38424309
3843 - `type: "coordinator"`
3844
3845 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
3846
3847 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
3848
3849 - `string`
3850
3851 - `BetaManagedAgentsAgentParams object`
3852
3853 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
3854
3855 - `type: "agent"`
3856
3857 - `id: string`
3858
3859 The `agent` ID.
4310 - `BetaManagedAgentsMultiagentCoordinatorParams object`
4311
4312 A coordinator topology: the session's primary thread orchestrates work by spawning session threads, each running an agent drawn from the `agents` roster.
4313
4314 - `type: "coordinator"`
4315
4316 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
4317
4318 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
4319
4320 - `string`
4321
4322 - `BetaManagedAgentsAgentParams object`
4323
4324 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
4325
4326 - `type: "agent"`
4327
4328 - `id: string`
4329
4330 The `agent` ID.
4331
4332 minLength: 1, maxLength: 128
4333
4334 - `version: optional number`
4335
4336 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
4337
4338 format: int32
4339
4340 - `BetaManagedAgentsMultiagentSelfParams object`
4341
4342 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
4343
4344 - `type: "self"`
4345
4346 - `BetaManagedAgentsAdvisorParams object`
4347
4348 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
4349
4350 - `type: "advisor"`
4351
4352 - `model: string`
4353
4354 A Claude model id. The model must be permitted as an advisor for this agent's model.
4355
4356 minLength: 1, maxLength: 256
4357
4358 - `BetaManagedAgentsMultiagent20261001Params object`
4359
4360 Multiagent configuration with three members, each enabled or disabled on its own. On an update, if the agent's stored `multiagent` also has type `multiagent_20261001`, this configuration is merged into the stored one, level by level, instead of replacing it. A key that the update omits keeps its stored value. A key sent as null takes its default, on create as well, so `"workflows": null` enables workflows. An object sent with a `type` other than the stored one replaces the stored object, and the keys that it omits take their defaults. A `predefined_agents` list that is sent replaces the stored list. Every object that is sent needs its `type`, and an enabled `advisor` needs its `model`. Other validation applies to the merged result.
4361
4362 - `type: "multiagent_20261001"`
4363
4364 - `advisor: optional BetaManagedAgentsMultiagentAdvisorParams or null`
4365
4366 Whether the session's primary thread can consult an advisor model. Defaults to disabled.
4367
4368 - `BetaManagedAgentsMultiagentAdvisorEnabledParams object`
4369
4370 The session's primary thread can consult `model` mid-turn.
4371
4372 - `type: "enabled"`
4373
4374 - `model: string`
4375
4376 A Claude model id. The model must be permitted as an advisor for this agent's model.
4377
4378 minLength: 1, maxLength: 256
4379
4380 - `BetaManagedAgentsMultiagentAdvisorDisabledParams object`
4381
4382 The agent has no advisor.
4383
4384 - `type: "disabled"`
4385
4386 - `subagents: optional BetaManagedAgentsMultiagentSubagentsParams or null`
4387
4388 Whether the agent can spawn session threads. Defaults to enabled.
4389
4390 - `BetaManagedAgentsMultiagentSubagentsEnabledParams object`
4391
4392 The agent can spawn session threads. Each thread runs a predefined agent, which is a saved agent in `predefined_agents`, or an inline agent, which the agent defines when it spawns the thread and which is not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
4393
4394 - `type: "enabled"`
4395
4396 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
4397
4398 Whether the agent can define inline agents when it spawns session threads. Defaults to enabled.
4399
4400 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
4401
4402 The agent can define inline agents.
4403
4404 - `type: "enabled"`
4405
4406 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
4407
4408 The agent cannot define inline agents.
4409
4410 - `type: "disabled"`
4411
4412 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
4413
4414 Predefined agents that this agent can spawn as session threads. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `workflows.predefined_agents`, and an agent in one list is not added to the other.
4415
4416 - `string`
4417
4418 - `BetaManagedAgentsAgentParams object`
4419
4420 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
4421
4422 - `BetaManagedAgentsMultiagentSelfParams object`
4423
4424 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
4425
4426 - `BetaManagedAgentsMultiagentSubagentsDisabledParams object`
4427
4428 The agent cannot spawn session threads.
4429
4430 - `type: "disabled"`
4431
4432 - `workflows: optional BetaManagedAgentsMultiagentWorkflowsParams or null`
4433
4434 Whether the agent can start workflow runs. Defaults to enabled.
4435
4436 - `BetaManagedAgentsMultiagentWorkflowsEnabledParams object`
4437
4438 The agent can start workflow runs. Each run follows a plan, a program that the agent writes. A plan can use predefined agents, which are the saved agents in `predefined_agents`, and inline agents, which it defines itself and which are not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
4439
4440 - `type: "enabled"`
4441
4442 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
4443
4444 Whether a run's plan can define inline agents. Defaults to enabled.
4445
4446 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
4447
4448 Predefined agents that a run's plan can use. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `subagents.predefined_agents`, and an agent in one list is not added to the other.
4449
4450 - `string`
4451
4452 - `BetaManagedAgentsAgentParams object`
4453
4454 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
4455
4456 - `BetaManagedAgentsMultiagentSelfParams object`
4457
4458 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
4459
4460 - `BetaManagedAgentsMultiagentWorkflowsDisabledParams object`
4461
4462 The agent cannot start workflow runs.
4463
4464 - `type: "disabled"`
4465
4466- `name: optional string`
4467
4468 Human-readable name. Must be non-empty. Omit to preserve. Cannot be cleared.
4469
4470 maxLength: 256
4471
4472- `skills: optional array of BetaManagedAgentsSkillParams or null`
4473
4474 Skills. Full replacement. Omit to preserve; send empty array or null to clear.
4475
4476 - `BetaManagedAgentsAnthropicSkillParams object`
4477
4478 An Anthropic-managed skill.
4479
4480 - `type: "anthropic"`
4481
4482 - `skill_id: string`
4483
4484 Identifier of the Anthropic skill (e.g., "xlsx").
4485
4486 minLength: 1, maxLength: 64
4487
4488 - `version: optional string or null`
4489
4490 Version to pin. Defaults to latest if omitted.
4491
4492 minLength: 1, maxLength: 64
4493
4494 - `BetaManagedAgentsCustomSkillParams object`
4495
4496 A user-created custom skill.
4497
4498 - `type: "custom"`
4499
4500 - `skill_id: string`
4501
4502 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
4503
4504 minLength: 1, maxLength: 64
4505
4506 - `version: optional string or null`
4507
4508 Version to pin. Defaults to latest if omitted.
4509
4510 minLength: 1, maxLength: 64
4511
4512- `system: optional string or null`
4513
4514 System prompt. Omit to preserve; send empty string or null to clear.
4515
4516 maxLength: 100000
4517
4518- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams or null`
4519
4520 Tool configurations available to the agent. Full replacement. Omit to preserve; send empty array or null to clear. Maximum of 256 tools across all toolsets allowed.
4521
4522 - `BetaManagedAgentsAgentToolset20260401Params object`
4523
4524 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
4525
4526 - `type: "agent_toolset_20260401"`
4527
4528 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
4529
4530 Per-tool configuration overrides.
4531
4532 - `BetaManagedAgentsBashToolConfigParams object`
4533
4534 Configuration override for the bash tool.
4535
4536 - `type: optional "bash"`
4537
4538 - `name: "bash"`
4539
4540 Must be "bash".
4541
4542 - `enabled: optional boolean or null`
4543
4544 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4545
4546 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4547
4548 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4549
4550 - `BetaManagedAgentsAlwaysAllowPolicy object`
4551
4552 Tool calls are automatically approved without user confirmation.
4553
4554 - `type: "always_allow"`
4555
4556 - `BetaManagedAgentsAlwaysAskPolicy object`
4557
4558 Tool calls require user confirmation before execution.
4559
4560 - `type: "always_ask"`
4561
4562 - `BetaManagedAgentsAutoPolicy object`
4563
4564 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4565
4566 - `type: "auto"`
4567
4568 - `BetaManagedAgentsEditToolConfigParams object`
4569
4570 Configuration override for the edit tool.
4571
4572 - `type: optional "edit"`
4573
4574 - `name: "edit"`
4575
4576 Must be "edit".
4577
4578 - `enabled: optional boolean or null`
4579
4580 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4581
4582 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4583
4584 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4585
4586 - `BetaManagedAgentsAlwaysAllowPolicy object`
4587
4588 Tool calls are automatically approved without user confirmation.
4589
4590 - `BetaManagedAgentsAlwaysAskPolicy object`
4591
4592 Tool calls require user confirmation before execution.
4593
4594 - `BetaManagedAgentsAutoPolicy object`
4595
4596 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4597
4598 - `BetaManagedAgentsReadToolConfigParams object`
4599
4600 Configuration override for the read tool.
4601
4602 - `type: optional "read"`
4603
4604 - `name: "read"`
4605
4606 Must be "read".
4607
4608 - `enabled: optional boolean or null`
4609
4610 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4611
4612 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4613
4614 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4615
4616 - `BetaManagedAgentsAlwaysAllowPolicy object`
4617
4618 Tool calls are automatically approved without user confirmation.
4619
4620 - `BetaManagedAgentsAlwaysAskPolicy object`
4621
4622 Tool calls require user confirmation before execution.
4623
4624 - `BetaManagedAgentsAutoPolicy object`
4625
4626 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4627
4628 - `BetaManagedAgentsWriteToolConfigParams object`
4629
4630 Configuration override for the write tool.
4631
4632 - `type: optional "write"`
4633
4634 - `name: "write"`
4635
4636 Must be "write".
4637
4638 - `enabled: optional boolean or null`
4639
4640 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4641
4642 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4643
4644 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4645
4646 - `BetaManagedAgentsAlwaysAllowPolicy object`
4647
4648 Tool calls are automatically approved without user confirmation.
4649
4650 - `BetaManagedAgentsAlwaysAskPolicy object`
4651
4652 Tool calls require user confirmation before execution.
4653
4654 - `BetaManagedAgentsAutoPolicy object`
4655
4656 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4657
4658 - `BetaManagedAgentsGlobToolConfigParams object`
4659
4660 Configuration override for the glob tool.
4661
4662 - `type: optional "glob"`
4663
4664 - `name: "glob"`
4665
4666 Must be "glob".
4667
4668 - `enabled: optional boolean or null`
4669
4670 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4671
4672 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4673
4674 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4675
4676 - `BetaManagedAgentsAlwaysAllowPolicy object`
4677
4678 Tool calls are automatically approved without user confirmation.
4679
4680 - `BetaManagedAgentsAlwaysAskPolicy object`
4681
4682 Tool calls require user confirmation before execution.
4683
4684 - `BetaManagedAgentsAutoPolicy object`
4685
4686 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4687
4688 - `BetaManagedAgentsGrepToolConfigParams object`
4689
4690 Configuration override for the grep tool.
4691
4692 - `type: optional "grep"`
4693
4694 - `name: "grep"`
4695
4696 Must be "grep".
4697
4698 - `enabled: optional boolean or null`
4699
4700 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4701
4702 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4703
4704 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4705
4706 - `BetaManagedAgentsAlwaysAllowPolicy object`
4707
4708 Tool calls are automatically approved without user confirmation.
4709
4710 - `BetaManagedAgentsAlwaysAskPolicy object`
4711
4712 Tool calls require user confirmation before execution.
4713
4714 - `BetaManagedAgentsAutoPolicy object`
4715
4716 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4717
4718 - `BetaManagedAgentsWebFetchToolConfigParams object`
4719
4720 Configuration override for the web_fetch tool.
4721
4722 - `type: optional "web_fetch"`
4723
4724 - `name: "web_fetch"`
4725
4726 Must be "web_fetch".
4727
4728 - `allowed_domains: optional array of string`
4729
4730 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
4731
4732 - `blocked_domains: optional array of string`
4733
4734 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
4735
4736 - `enabled: optional boolean or null`
4737
4738 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4739
4740 - `max_content_tokens: optional number or null`
4741
4742 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
4743
4744 format: int32
4745
4746 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4747
4748 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4749
4750 - `BetaManagedAgentsAlwaysAllowPolicy object`
4751
4752 Tool calls are automatically approved without user confirmation.
4753
4754 - `BetaManagedAgentsAlwaysAskPolicy object`
4755
4756 Tool calls require user confirmation before execution.
4757
4758 - `BetaManagedAgentsAutoPolicy object`
4759
4760 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4761
4762 - `url_sources: optional BetaManagedAgentsWebFetchURLSourcesParams or null`
4763
4764 Which sources contribute URLs the tool may fetch. Omit to allow every source.
4765
4766 - `client_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
4767
4768 Which custom tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be a custom tool in the same tools array.
4769
4770 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
4771
4772 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
4773
4774 - `"all"`
4775
4776 - `"none"`
4777
4778 - `BetaManagedAgentsWebFetchURLSourceToolFilter = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone or BetaManagedAgentsWebFetchURLSourceOnly or BetaManagedAgentsWebFetchURLSourceExcept`
4779
4780 Which tools' results contribute URLs that may be fetched.
4781
4782 - `BetaManagedAgentsWebFetchURLSourceAll object`
4783
4784 Every URL from this source may be fetched. This is the default.
4785
4786 - `type: "all"`
4787
4788 - `BetaManagedAgentsWebFetchURLSourceNone object`
4789
4790 This source contributes no URLs that may be fetched.
4791
4792 - `type: "none"`
4793
4794 - `BetaManagedAgentsWebFetchURLSourceOnly object`
4795
4796 Only the named tools' results contribute URLs that may be fetched.
4797
4798 - `type: "only"`
4799
4800 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
4801
4802 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
4803
4804 - `type: "tool_reference"`
4805
4806 Must be "tool_reference".
4807
4808 - `name: string`
4809
4810 Name of the tool. Compared exactly, so upper and lower case letters are different.
4811
4812 minLength: 1, maxLength: 128
4813
4814 - `BetaManagedAgentsWebFetchURLSourceExcept object`
4815
4816 Every tool's results contribute URLs that may be fetched, except the named tools' results.
4817
4818 - `type: "except"`
4819
4820 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
4821
4822 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
4823
4824 - `type: "tool_reference"`
4825
4826 Must be "tool_reference".
4827
4828 - `name: string`
4829
4830 Name of the tool. Compared exactly, so upper and lower case letters are different.
4831
4832 minLength: 1, maxLength: 128
4833
4834 - `server_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
4835
4836 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be "web_search" or "web_fetch".
4837
4838 - `user_input: optional BetaManagedAgentsWebFetchURLSourceUserInputParams or null`
4839
4840 Whether URLs in the text of user messages may be fetched: "all" (the default) or "none".
4841
4842 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
4843
4844 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
4845
4846 - `BetaManagedAgentsWebFetchURLSourceUserInput = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone`
4847
4848 Whether URLs in the text of user messages may be fetched.
4849
4850 - `BetaManagedAgentsWebFetchURLSourceAll object`
4851
4852 Every URL from this source may be fetched. This is the default.
4853
4854 - `BetaManagedAgentsWebFetchURLSourceNone object`
4855
4856 This source contributes no URLs that may be fetched.
4857
4858 - `BetaManagedAgentsWebSearchToolConfigParams object`
4859
4860 Configuration override for the web_search tool.
4861
4862 - `type: optional "web_search"`
4863
4864 - `name: "web_search"`
4865
4866 Must be "web_search".
4867
4868 - `allowed_domains: optional array of string`
4869
4870 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
4871
4872 - `blocked_domains: optional array of string`
4873
4874 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
4875
4876 - `enabled: optional boolean or null`
4877
4878 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4879
4880 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4881
4882 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4883
4884 - `BetaManagedAgentsAlwaysAllowPolicy object`
4885
4886 Tool calls are automatically approved without user confirmation.
4887
4888 - `BetaManagedAgentsAlwaysAskPolicy object`
4889
4890 Tool calls require user confirmation before execution.
4891
4892 - `BetaManagedAgentsAutoPolicy object`
4893
4894 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4895
4896 - `user_location: optional BetaManagedAgentsUserLocation or null`
4897
4898 Approximate user location for search result localization.
4899
4900 - `type: "approximate"`
4901
4902 Location precision. Only "approximate" is supported.
4903
4904 - `city: optional string or null`
4905
4906 City name.
4907
4908 minLength: 1, maxLength: 255
4909
4910 - `country: optional string or null`
4911
4912 Two-letter ISO 3166-1 country code, uppercase.
4913
4914 - `region: optional string or null`
4915
4916 Region or state name.
4917
4918 minLength: 1, maxLength: 255
4919
4920 - `timezone: optional string or null`
4921
4922 IANA timezone identifier, e.g. "America/Los_Angeles".
4923
4924 minLength: 1, maxLength: 255
4925
4926 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
4927
4928 Default configuration applied to all tools in this set.
4929
4930 - `enabled: optional boolean or null`
4931
4932 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
4933
4934 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4935
4936 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
4937
4938 - `BetaManagedAgentsAlwaysAllowPolicy object`
4939
4940 Tool calls are automatically approved without user confirmation.
4941
4942 - `BetaManagedAgentsAlwaysAskPolicy object`
4943
4944 Tool calls require user confirmation before execution.
4945
4946 - `BetaManagedAgentsAutoPolicy object`
4947
4948 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4949
4950 - `BetaManagedAgentsMCPToolsetParams object`
4951
4952 Configuration for tools from an MCP server defined in `mcp_servers`.
4953
4954 - `type: "mcp_toolset"`
4955
4956 - `mcp_server_name: string`
4957
4958 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
4959
4960 minLength: 1, maxLength: 255
4961
4962 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
4963
4964 Per-tool configuration overrides.
4965
4966 - `name: string`
4967
4968 Name of the MCP tool to configure. 1-128 characters.
38604969
38614970 minLength: 1, maxLength: 128
38624971
3863 - `version: optional number`
3864
3865 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
3866
3867 format: int32
3868
3869 - `BetaManagedAgentsMultiagentSelfParams object`
3870
3871 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
3872
3873 - `type: "self"`
3874
3875 - `BetaManagedAgentsAdvisorParams object`
3876
3877 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
3878
3879 - `type: "advisor"`
3880
3881 - `model: string`
3882
3883 A Claude model id. The model must be permitted as an advisor for this agent's model.
3884
3885 minLength: 1, maxLength: 256
3886
3887- `name: optional string`
3888
3889 Human-readable name. Must be non-empty. Omit to preserve. Cannot be cleared.
3890
3891 maxLength: 256
3892
3893- `skills: optional array of BetaManagedAgentsSkillParams or null`
3894
3895 Skills. Full replacement. Omit to preserve; send empty array or null to clear.
3896
3897 - `BetaManagedAgentsAnthropicSkillParams object`
3898
3899 An Anthropic-managed skill.
3900
3901 - `type: "anthropic"`
3902
3903 - `skill_id: string`
3904
3905 Identifier of the Anthropic skill (e.g., "xlsx").
3906
3907 minLength: 1, maxLength: 64
3908
3909 - `version: optional string or null`
3910
3911 Version to pin. Defaults to latest if omitted.
3912
3913 minLength: 1, maxLength: 64
3914
3915 - `BetaManagedAgentsCustomSkillParams object`
3916
3917 A user-created custom skill.
4972 - `enabled: optional boolean or null`
4973
4974 Whether this tool is enabled. Overrides the `default_config` setting.
4975
4976 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4977
4978 Permission policy for this tool. Overrides the `default_config` setting.
4979
4980 - `BetaManagedAgentsAlwaysAllowPolicy object`
4981
4982 Tool calls are automatically approved without user confirmation.
4983
4984 - `BetaManagedAgentsAlwaysAskPolicy object`
4985
4986 Tool calls require user confirmation before execution.
4987
4988 - `BetaManagedAgentsAutoPolicy object`
4989
4990 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4991
4992 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
4993
4994 Default configuration for all tools from this server.
4995
4996 - `enabled: optional boolean or null`
4997
4998 Whether tools are enabled by default. Defaults to true if not specified.
4999
5000 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
5001
5002 Default permission policy for tools from this server.
5003
5004 - `BetaManagedAgentsAlwaysAllowPolicy object`
5005
5006 Tool calls are automatically approved without user confirmation.
5007
5008 - `BetaManagedAgentsAlwaysAskPolicy object`
5009
5010 Tool calls require user confirmation before execution.
5011
5012 - `BetaManagedAgentsAutoPolicy object`
5013
5014 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
5015
5016 - `BetaManagedAgentsCustomToolParams object`
5017
5018 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
39185019
39195020 - `type: "custom"`
39205021
3921 - `skill_id: string`
3922
3923 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
3924
3925 minLength: 1, maxLength: 64
3926
3927 - `version: optional string or null`
3928
3929 Version to pin. Defaults to latest if omitted.
3930
3931 minLength: 1, maxLength: 64
3932
3933- `system: optional string or null`
3934
3935 System prompt. Omit to preserve; send empty string or null to clear.
3936
3937 maxLength: 100000
3938
3939- `tools: optional array of BetaManagedAgentsAgentToolset20260401Params or BetaManagedAgentsMCPToolsetParams or BetaManagedAgentsCustomToolParams or null`
3940
3941 Tool configurations available to the agent. Full replacement. Omit to preserve; send empty array or null to clear. Maximum of 256 tools across all toolsets allowed.
3942
3943 - `BetaManagedAgentsAgentToolset20260401Params object`
3944
3945 Configuration for built-in agent tools. Use this to enable or disable groups of tools available to the agent.
3946
3947 - `type: "agent_toolset_20260401"`
3948
3949 - `configs: optional array of BetaManagedAgentsAgentToolConfigParams`
3950
3951 Per-tool configuration overrides.
3952
3953 - `BetaManagedAgentsBashToolConfigParams object`
3954
3955 Configuration override for the bash tool.
3956
3957 - `type: optional "bash"`
3958
3959 - `name: "bash"`
3960
3961 Must be "bash".
3962
3963 - `enabled: optional boolean or null`
3964
3965 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
3966
3967 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
3968
3969 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
3970
3971 - `BetaManagedAgentsAlwaysAllowPolicy object`
3972
3973 Tool calls are automatically approved without user confirmation.
3974
3975 - `type: "always_allow"`
3976
3977 - `BetaManagedAgentsAlwaysAskPolicy object`
3978
3979 Tool calls require user confirmation before execution.
3980
3981 - `type: "always_ask"`
3982
3983 - `BetaManagedAgentsAutoPolicy object`
3984
3985 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
3986
3987 - `type: "auto"`
3988
3989 - `BetaManagedAgentsEditToolConfigParams object`
3990
3991 Configuration override for the edit tool.
3992
3993 - `type: optional "edit"`
3994
3995 - `name: "edit"`
3996
3997 Must be "edit".
3998
3999 - `enabled: optional boolean or null`
4000
4001 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4002
4003 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4004
4005 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4006
4007 - `BetaManagedAgentsAlwaysAllowPolicy object`
4008
4009 Tool calls are automatically approved without user confirmation.
4010
4011 - `BetaManagedAgentsAlwaysAskPolicy object`
4012
4013 Tool calls require user confirmation before execution.
4014
4015 - `BetaManagedAgentsAutoPolicy object`
4016
4017 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4018
4019 - `BetaManagedAgentsReadToolConfigParams object`
4020
4021 Configuration override for the read tool.
4022
4023 - `type: optional "read"`
4024
4025 - `name: "read"`
4026
4027 Must be "read".
4028
4029 - `enabled: optional boolean or null`
4030
4031 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4032
4033 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4034
4035 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4036
4037 - `BetaManagedAgentsAlwaysAllowPolicy object`
4038
4039 Tool calls are automatically approved without user confirmation.
4040
4041 - `BetaManagedAgentsAlwaysAskPolicy object`
4042
4043 Tool calls require user confirmation before execution.
4044
4045 - `BetaManagedAgentsAutoPolicy object`
4046
4047 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4048
4049 - `BetaManagedAgentsWriteToolConfigParams object`
4050
4051 Configuration override for the write tool.
4052
4053 - `type: optional "write"`
4054
4055 - `name: "write"`
4056
4057 Must be "write".
4058
4059 - `enabled: optional boolean or null`
4060
4061 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4062
4063 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4064
4065 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4066
4067 - `BetaManagedAgentsAlwaysAllowPolicy object`
4068
4069 Tool calls are automatically approved without user confirmation.
4070
4071 - `BetaManagedAgentsAlwaysAskPolicy object`
4072
4073 Tool calls require user confirmation before execution.
4074
4075 - `BetaManagedAgentsAutoPolicy object`
4076
4077 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4078
4079 - `BetaManagedAgentsGlobToolConfigParams object`
4080
4081 Configuration override for the glob tool.
4082
4083 - `type: optional "glob"`
4084
4085 - `name: "glob"`
4086
4087 Must be "glob".
4088
4089 - `enabled: optional boolean or null`
4090
4091 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4092
4093 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4094
4095 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4096
4097 - `BetaManagedAgentsAlwaysAllowPolicy object`
4098
4099 Tool calls are automatically approved without user confirmation.
4100
4101 - `BetaManagedAgentsAlwaysAskPolicy object`
4102
4103 Tool calls require user confirmation before execution.
4104
4105 - `BetaManagedAgentsAutoPolicy object`
4106
4107 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4108
4109 - `BetaManagedAgentsGrepToolConfigParams object`
4110
4111 Configuration override for the grep tool.
4112
4113 - `type: optional "grep"`
4114
4115 - `name: "grep"`
4116
4117 Must be "grep".
4118
4119 - `enabled: optional boolean or null`
4120
4121 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4122
4123 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4124
4125 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4126
4127 - `BetaManagedAgentsAlwaysAllowPolicy object`
4128
4129 Tool calls are automatically approved without user confirmation.
4130
4131 - `BetaManagedAgentsAlwaysAskPolicy object`
4132
4133 Tool calls require user confirmation before execution.
4134
4135 - `BetaManagedAgentsAutoPolicy object`
4136
4137 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4138
4139 - `BetaManagedAgentsWebFetchToolConfigParams object`
4140
4141 Configuration override for the web_fetch tool.
4142
4143 - `type: optional "web_fetch"`
4144
4145 - `name: "web_fetch"`
4146
4147 Must be "web_fetch".
4148
4149 - `allowed_domains: optional array of string`
4150
4151 Only fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
4152
4153 - `blocked_domains: optional array of string`
4154
4155 Never fetch URLs whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme, port, or path). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
4156
4157 - `enabled: optional boolean or null`
4158
4159 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4160
4161 - `max_content_tokens: optional number or null`
4162
4163 Maximum number of tokens of fetched text content to include in context per call. Does not apply to binary content such as PDFs.
4164
4165 format: int32
4166
4167 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4168
4169 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4170
4171 - `BetaManagedAgentsAlwaysAllowPolicy object`
4172
4173 Tool calls are automatically approved without user confirmation.
4174
4175 - `BetaManagedAgentsAlwaysAskPolicy object`
4176
4177 Tool calls require user confirmation before execution.
4178
4179 - `BetaManagedAgentsAutoPolicy object`
4180
4181 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4182
4183 - `url_sources: optional BetaManagedAgentsWebFetchURLSourcesParams or null`
4184
4185 Which sources contribute URLs the tool may fetch. Omit to allow every source.
4186
4187 - `client_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
4188
4189 Which custom tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be a custom tool in the same tools array.
4190
4191 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
4192
4193 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
4194
4195 - `"all"`
4196
4197 - `"none"`
4198
4199 - `BetaManagedAgentsWebFetchURLSourceToolFilter = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone or BetaManagedAgentsWebFetchURLSourceOnly or BetaManagedAgentsWebFetchURLSourceExcept`
4200
4201 Which tools' results contribute URLs that may be fetched.
4202
4203 - `BetaManagedAgentsWebFetchURLSourceAll object`
4204
4205 Every URL from this source may be fetched. This is the default.
4206
4207 - `type: "all"`
4208
4209 - `BetaManagedAgentsWebFetchURLSourceNone object`
4210
4211 This source contributes no URLs that may be fetched.
4212
4213 - `type: "none"`
4214
4215 - `BetaManagedAgentsWebFetchURLSourceOnly object`
4216
4217 Only the named tools' results contribute URLs that may be fetched.
4218
4219 - `type: "only"`
4220
4221 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
4222
4223 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
4224
4225 - `type: "tool_reference"`
4226
4227 Must be "tool_reference".
4228
4229 - `name: string`
4230
4231 Name of the tool. Compared exactly, so upper and lower case letters are different.
4232
4233 minLength: 1, maxLength: 128
4234
4235 - `BetaManagedAgentsWebFetchURLSourceExcept object`
4236
4237 Every tool's results contribute URLs that may be fetched, except the named tools' results.
4238
4239 - `type: "except"`
4240
4241 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
4242
4243 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
4244
4245 - `type: "tool_reference"`
4246
4247 Must be "tool_reference".
4248
4249 - `name: string`
4250
4251 Name of the tool. Compared exactly, so upper and lower case letters are different.
4252
4253 minLength: 1, maxLength: 128
4254
4255 - `server_tool_results: optional BetaManagedAgentsWebFetchURLSourceToolFilterParams or null`
4256
4257 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched: "all" (the default), "none", or an only or except list. Each name in a list must be "web_search" or "web_fetch".
4258
4259 - `user_input: optional BetaManagedAgentsWebFetchURLSourceUserInputParams or null`
4260
4261 Whether URLs in the text of user messages may be fetched: "all" (the default) or "none".
4262
4263 - `BetaManagedAgentsWebFetchURLSourceShorthand = "all" or "none"`
4264
4265 String form of a url_sources value that has no field other than its type: "all" means {"type": "all"} and "none" means {"type": "none"}.
4266
4267 - `BetaManagedAgentsWebFetchURLSourceUserInput = BetaManagedAgentsWebFetchURLSourceAll or BetaManagedAgentsWebFetchURLSourceNone`
4268
4269 Whether URLs in the text of user messages may be fetched.
4270
4271 - `BetaManagedAgentsWebFetchURLSourceAll object`
4272
4273 Every URL from this source may be fetched. This is the default.
4274
4275 - `BetaManagedAgentsWebFetchURLSourceNone object`
4276
4277 This source contributes no URLs that may be fetched.
4278
4279 - `BetaManagedAgentsWebSearchToolConfigParams object`
4280
4281 Configuration override for the web_search tool.
4282
4283 - `type: optional "web_search"`
4284
4285 - `name: "web_search"`
4286
4287 Must be "web_search".
4288
4289 - `allowed_domains: optional array of string`
4290
4291 Only return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "docs.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with blocked_domains.
4292
4293 - `blocked_domains: optional array of string`
4294
4295 Never return search results whose host is one of these domains or a subdomain of one. Each entry is a plain hostname like "ads.example.com" (no scheme or port; an optional path suffix is accepted). At most 64 entries; an empty list is rejected (omit the field instead). Cannot be combined with allowed_domains.
4296
4297 - `enabled: optional boolean or null`
4298
4299 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
4300
4301 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4302
4303 Permission policy for this tool. Controls whether tool calls are auto-approved or require confirmation.
4304
4305 - `BetaManagedAgentsAlwaysAllowPolicy object`
4306
4307 Tool calls are automatically approved without user confirmation.
4308
4309 - `BetaManagedAgentsAlwaysAskPolicy object`
4310
4311 Tool calls require user confirmation before execution.
4312
4313 - `BetaManagedAgentsAutoPolicy object`
4314
4315 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4316
4317 - `user_location: optional BetaManagedAgentsUserLocation or null`
4318
4319 Approximate user location for search result localization.
4320
4321 - `type: "approximate"`
4322
4323 Location precision. Only "approximate" is supported.
4324
4325 - `city: optional string or null`
4326
4327 City name.
4328
4329 minLength: 1, maxLength: 255
4330
4331 - `country: optional string or null`
4332
4333 Two-letter ISO 3166-1 country code, uppercase.
4334
4335 - `region: optional string or null`
4336
4337 Region or state name.
4338
4339 minLength: 1, maxLength: 255
4340
4341 - `timezone: optional string or null`
4342
4343 IANA timezone identifier, e.g. "America/Los_Angeles".
4344
4345 minLength: 1, maxLength: 255
4346
4347 - `default_config: optional BetaManagedAgentsAgentToolsetDefaultConfigParams or null`
4348
4349 Default configuration applied to all tools in this set.
4350
4351 - `enabled: optional boolean or null`
4352
4353 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
4354
4355 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4356
4357 Default permission policy for tools. Controls whether tool calls are auto-approved or require confirmation.
4358
4359 - `BetaManagedAgentsAlwaysAllowPolicy object`
4360
4361 Tool calls are automatically approved without user confirmation.
4362
4363 - `BetaManagedAgentsAlwaysAskPolicy object`
4364
4365 Tool calls require user confirmation before execution.
4366
4367 - `BetaManagedAgentsAutoPolicy object`
4368
4369 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4370
4371 - `BetaManagedAgentsMCPToolsetParams object`
4372
4373 Configuration for tools from an MCP server defined in `mcp_servers`.
4374
4375 - `type: "mcp_toolset"`
4376
4377 - `mcp_server_name: string`
4378
4379 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
4380
4381 minLength: 1, maxLength: 255
4382
4383 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
4384
4385 Per-tool configuration overrides.
4386
4387 - `name: string`
4388
4389 Name of the MCP tool to configure. 1-128 characters.
4390
4391 minLength: 1, maxLength: 128
4392
4393 - `enabled: optional boolean or null`
4394
4395 Whether this tool is enabled. Overrides the `default_config` setting.
4396
4397 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4398
4399 Permission policy for this tool. Overrides the `default_config` setting.
4400
4401 - `BetaManagedAgentsAlwaysAllowPolicy object`
4402
4403 Tool calls are automatically approved without user confirmation.
4404
4405 - `BetaManagedAgentsAlwaysAskPolicy object`
4406
4407 Tool calls require user confirmation before execution.
4408
4409 - `BetaManagedAgentsAutoPolicy object`
4410
4411 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4412
4413 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
4414
4415 Default configuration for all tools from this server.
4416
4417 - `enabled: optional boolean or null`
4418
4419 Whether tools are enabled by default. Defaults to true if not specified.
4420
4421 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
4422
4423 Default permission policy for tools from this server.
4424
4425 - `BetaManagedAgentsAlwaysAllowPolicy object`
4426
4427 Tool calls are automatically approved without user confirmation.
4428
4429 - `BetaManagedAgentsAlwaysAskPolicy object`
4430
4431 Tool calls require user confirmation before execution.
4432
4433 - `BetaManagedAgentsAutoPolicy object`
4434
4435 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
4436
4437 - `BetaManagedAgentsCustomToolParams object`
4438
4439 A custom tool that is executed by the API client rather than the agent. When the agent calls this tool, an `agent.custom_tool_use` event is emitted and the session goes idle, waiting for the client to provide the result via a `user.custom_tool_result` event.
4440
4441 - `type: "custom"`
4442
44435022 - `description: string`
44445023
44455024 Description of what the tool does, shown to the agent to help it decide when to use the tool.
from line 5137
45585137
45595138 - `"claude-haiku-4-5"`
45605139
4561 Fastest model with near-frontier intelligence
4562
45635140 - `"claude-haiku-4-5-20251001"`
45645141
4565 Fastest model with near-frontier intelligence
4566
45675142 - `"claude-opus-4-5"`
45685143
45695144 Powerful intelligence for long-running agents and coding
from line 5211
46365211
46375212 Multiagent orchestration configuration. Null when the agent is single-threaded.
46385213
4639 - `type: "coordinator"`
4640
4641 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
4642
4643 Agents the coordinator may spawn as session threads, each resolved to a specific version.
4644
4645 - `BetaManagedAgentsAgentReference object`
4646
4647 A resolved agent reference with a concrete version.
4648
4649 - `type: "agent"`
4650
4651 - `id: string`
4652
4653 - `version: number`
4654
4655 format: int32
4656
4657 - `BetaManagedAgentsAdvisor object`
4658
4659 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
4660
4661 - `type: "advisor"`
4662
4663 - `model: string`
4664
4665 The advisor model id.
5214 - `BetaManagedAgentsMultiagentCoordinator object`
5215
5216 Resolved coordinator topology with a concrete agent roster.
5217
5218 - `type: "coordinator"`
5219
5220 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
5221
5222 Agents the coordinator may spawn as session threads, each resolved to a specific version.
5223
5224 - `BetaManagedAgentsAgentReference object`
5225
5226 A resolved agent reference with a concrete version.
5227
5228 - `type: "agent"`
5229
5230 - `id: string`
5231
5232 - `version: number`
5233
5234 format: int32
5235
5236 - `BetaManagedAgentsAdvisor object`
5237
5238 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
5239
5240 - `type: "advisor"`
5241
5242 - `model: string`
5243
5244 The advisor model id.
5245
5246 - `BetaManagedAgentsMultiagent20261001 object`
5247
5248 Resolved multiagent configuration with three members, each enabled or disabled on its own.
5249
5250 - `type: "multiagent_20261001"`
5251
5252 - `advisor: BetaManagedAgentsMultiagentAdvisor`
5253
5254 Whether the session's primary thread can consult an advisor model.
5255
5256 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
5257
5258 The session's primary thread can consult `model` mid-turn.
5259
5260 - `type: "enabled"`
5261
5262 - `model: string`
5263
5264 The advisor model id.
5265
5266 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
5267
5268 The agent has no advisor.
5269
5270 - `type: "disabled"`
5271
5272 - `subagents: BetaManagedAgentsMultiagentSubagents`
5273
5274 Whether the agent can spawn session threads.
5275
5276 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
5277
5278 The agent can spawn session threads.
5279
5280 - `type: "enabled"`
5281
5282 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
5283
5284 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
5285
5286 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
5287
5288 The agent can define inline agents.
5289
5290 - `type: "enabled"`
5291
5292 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
5293
5294 The agent cannot define inline agents.
5295
5296 - `type: "disabled"`
5297
5298 - `predefined_agents: array of BetaManagedAgentsAgentReference`
5299
5300 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
5301
5302 - `type: "agent"`
5303
5304 - `id: string`
5305
5306 - `version: number`
5307
5308 format: int32
5309
5310 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
5311
5312 The agent cannot spawn session threads.
5313
5314 - `type: "disabled"`
5315
5316 - `workflows: BetaManagedAgentsMultiagentWorkflows`
5317
5318 Whether the agent can start workflow runs.
5319
5320 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
5321
5322 The agent can start workflow runs.
5323
5324 - `type: "enabled"`
5325
5326 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
5327
5328 Whether a run's plan can define inline agents, which are not saved.
5329
5330 - `predefined_agents: array of BetaManagedAgentsAgentReference`
5331
5332 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
5333
5334 - `type: "agent"`
5335
5336 - `id: string`
5337
5338 - `version: number`
5339
5340 format: int32
5341
5342 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
5343
5344 The agent cannot start workflow runs.
5345
5346 - `type: "disabled"`
46665347
46675348 - `name: string`
46685349
from line 5825
51445825 -H "X-Api-Key: $ANTHROPIC_API_KEY" \
51455826 -d '{
51465827 "description": "updated",
5147 "multiagent": {
5148 "agents": [
5149 "agent_011CZkYqphY8vELVzwCUpqiQ",
5150 {
5151 "type": "self"
5152 }
5153 ],
5154 "type": "coordinator"
5155 },
51565828 "system": "You are a general-purpose agent that can research, write code, run commands, and use connected tools to complete the user'\''s task end to end.",
51575829 "version": 1
51585830 }'
from line 5857
51855857 "speed": "standard"
51865858 },
51875859 "multiagent": {
5188 "agents": [
5189 {
5190 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
5191 "type": "agent",
5192 "version": 1
5193 }
5194 ],
5195 "type": "coordinator"
5860 "advisor": {
5861 "type": "disabled"
5862 },
5863 "subagents": {
5864 "inline_agents": {
5865 "type": "enabled"
5866 },
5867 "predefined_agents": [
5868 {
5869 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
5870 "type": "agent",
5871 "version": 1
5872 }
5873 ],
5874 "type": "enabled"
5875 },
5876 "type": "multiagent_20261001",
5877 "workflows": {
5878 "inline_agents": {
5879 "type": "enabled"
5880 },
5881 "predefined_agents": [
5882 {
5883 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
5884 "type": "agent",
5885 "version": 1
5886 }
5887 ],
5888 "type": "enabled"
5889 }
51965890 },
51975891 "name": "My First Agent",
51985892 "skills": [
from line 6147
54536147
54546148 - `"claude-haiku-4-5"`
54556149
5456 Fastest model with near-frontier intelligence
5457
54586150 - `"claude-haiku-4-5-20251001"`
54596151
5460 Fastest model with near-frontier intelligence
5461
54626152 - `"claude-opus-4-5"`
54636153
54646154 Powerful intelligence for long-running agents and coding
from line 6221
55316221
55326222 Multiagent orchestration configuration. Null when the agent is single-threaded.
55336223
5534 - `type: "coordinator"`
5535
5536 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
5537
5538 Agents the coordinator may spawn as session threads, each resolved to a specific version.
5539
5540 - `BetaManagedAgentsAgentReference object`
5541
5542 A resolved agent reference with a concrete version.
5543
5544 - `type: "agent"`
5545
5546 - `id: string`
5547
5548 - `version: number`
5549
5550 format: int32
5551
5552 - `BetaManagedAgentsAdvisor object`
5553
5554 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
5555
5556 - `type: "advisor"`
5557
5558 - `model: string`
5559
5560 The advisor model id.
6224 - `BetaManagedAgentsMultiagentCoordinator object`
6225
6226 Resolved coordinator topology with a concrete agent roster.
6227
6228 - `type: "coordinator"`
6229
6230 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
6231
6232 Agents the coordinator may spawn as session threads, each resolved to a specific version.
6233
6234 - `BetaManagedAgentsAgentReference object`
6235
6236 A resolved agent reference with a concrete version.
6237
6238 - `type: "agent"`
6239
6240 - `id: string`
6241
6242 - `version: number`
6243
6244 format: int32
6245
6246 - `BetaManagedAgentsAdvisor object`
6247
6248 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
6249
6250 - `type: "advisor"`
6251
6252 - `model: string`
6253
6254 The advisor model id.
6255
6256 - `BetaManagedAgentsMultiagent20261001 object`
6257
6258 Resolved multiagent configuration with three members, each enabled or disabled on its own.
6259
6260 - `type: "multiagent_20261001"`
6261
6262 - `advisor: BetaManagedAgentsMultiagentAdvisor`
6263
6264 Whether the session's primary thread can consult an advisor model.
6265
6266 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
6267
6268 The session's primary thread can consult `model` mid-turn.
6269
6270 - `type: "enabled"`
6271
6272 - `model: string`
6273
6274 The advisor model id.
6275
6276 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
6277
6278 The agent has no advisor.
6279
6280 - `type: "disabled"`
6281
6282 - `subagents: BetaManagedAgentsMultiagentSubagents`
6283
6284 Whether the agent can spawn session threads.
6285
6286 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
6287
6288 The agent can spawn session threads.
6289
6290 - `type: "enabled"`
6291
6292 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
6293
6294 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
6295
6296 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
6297
6298 The agent can define inline agents.
6299
6300 - `type: "enabled"`
6301
6302 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
6303
6304 The agent cannot define inline agents.
6305
6306 - `type: "disabled"`
6307
6308 - `predefined_agents: array of BetaManagedAgentsAgentReference`
6309
6310 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
6311
6312 - `type: "agent"`
6313
6314 - `id: string`
6315
6316 - `version: number`
6317
6318 format: int32
6319
6320 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
6321
6322 The agent cannot spawn session threads.
6323
6324 - `type: "disabled"`
6325
6326 - `workflows: BetaManagedAgentsMultiagentWorkflows`
6327
6328 Whether the agent can start workflow runs.
6329
6330 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
6331
6332 The agent can start workflow runs.
6333
6334 - `type: "enabled"`
6335
6336 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
6337
6338 Whether a run's plan can define inline agents, which are not saved.
6339
6340 - `predefined_agents: array of BetaManagedAgentsAgentReference`
6341
6342 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
6343
6344 - `type: "agent"`
6345
6346 - `id: string`
6347
6348 - `version: number`
6349
6350 format: int32
6351
6352 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
6353
6354 The agent cannot start workflow runs.
6355
6356 - `type: "disabled"`
55616357
55626358 - `name: string`
55636359
from line 6862
60666862 "speed": "standard"
60676863 },
60686864 "multiagent": {
6069 "agents": [
6070 {
6071 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
6072 "type": "agent",
6073 "version": 1
6074 }
6075 ],
6076 "type": "coordinator"
6865 "advisor": {
6866 "type": "disabled"
6867 },
6868 "subagents": {
6869 "inline_agents": {
6870 "type": "enabled"
6871 },
6872 "predefined_agents": [
6873 {
6874 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
6875 "type": "agent",
6876 "version": 1
6877 }
6878 ],
6879 "type": "enabled"
6880 },
6881 "type": "multiagent_20261001",
6882 "workflows": {
6883 "inline_agents": {
6884 "type": "enabled"
6885 },
6886 "predefined_agents": [
6887 {
6888 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
6889 "type": "agent",
6890 "version": 1
6891 }
6892 ],
6893 "type": "enabled"
6894 }
60776895 },
60786896 "name": "My First Agent",
60796897 "skills": [
from line 7038
62207038
62217039 - `"claude-haiku-4-5"`
62227040
6223 Fastest model with near-frontier intelligence
6224
62257041 - `"claude-haiku-4-5-20251001"`
62267042
6227 Fastest model with near-frontier intelligence
6228
62297043 - `"claude-opus-4-5"`
62307044
62317045 Powerful intelligence for long-running agents and coding
from line 7112
62987112
62997113 Multiagent orchestration configuration. Null when the agent is single-threaded.
63007114
6301 - `type: "coordinator"`
6302
6303 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
6304
6305 Agents the coordinator may spawn as session threads, each resolved to a specific version.
6306
6307 - `BetaManagedAgentsAgentReference object`
6308
6309 A resolved agent reference with a concrete version.
6310
6311 - `type: "agent"`
6312
6313 - `id: string`
6314
6315 - `version: number`
6316
6317 format: int32
6318
6319 - `BetaManagedAgentsAdvisor object`
6320
6321 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
6322
6323 - `type: "advisor"`
6324
6325 - `model: string`
6326
6327 The advisor model id.
7115 - `BetaManagedAgentsMultiagentCoordinator object`
7116
7117 Resolved coordinator topology with a concrete agent roster.
7118
7119 - `type: "coordinator"`
7120
7121 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
7122
7123 Agents the coordinator may spawn as session threads, each resolved to a specific version.
7124
7125 - `BetaManagedAgentsAgentReference object`
7126
7127 A resolved agent reference with a concrete version.
7128
7129 - `type: "agent"`
7130
7131 - `id: string`
7132
7133 - `version: number`
7134
7135 format: int32
7136
7137 - `BetaManagedAgentsAdvisor object`
7138
7139 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
7140
7141 - `type: "advisor"`
7142
7143 - `model: string`
7144
7145 The advisor model id.
7146
7147 - `BetaManagedAgentsMultiagent20261001 object`
7148
7149 Resolved multiagent configuration with three members, each enabled or disabled on its own.
7150
7151 - `type: "multiagent_20261001"`
7152
7153 - `advisor: BetaManagedAgentsMultiagentAdvisor`
7154
7155 Whether the session's primary thread can consult an advisor model.
7156
7157 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
7158
7159 The session's primary thread can consult `model` mid-turn.
7160
7161 - `type: "enabled"`
7162
7163 - `model: string`
7164
7165 The advisor model id.
7166
7167 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
7168
7169 The agent has no advisor.
7170
7171 - `type: "disabled"`
7172
7173 - `subagents: BetaManagedAgentsMultiagentSubagents`
7174
7175 Whether the agent can spawn session threads.
7176
7177 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
7178
7179 The agent can spawn session threads.
7180
7181 - `type: "enabled"`
7182
7183 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
7184
7185 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
7186
7187 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
7188
7189 The agent can define inline agents.
7190
7191 - `type: "enabled"`
7192
7193 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
7194
7195 The agent cannot define inline agents.
7196
7197 - `type: "disabled"`
7198
7199 - `predefined_agents: array of BetaManagedAgentsAgentReference`
7200
7201 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
7202
7203 - `type: "agent"`
7204
7205 - `id: string`
7206
7207 - `version: number`
7208
7209 format: int32
7210
7211 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
7212
7213 The agent cannot spawn session threads.
7214
7215 - `type: "disabled"`
7216
7217 - `workflows: BetaManagedAgentsMultiagentWorkflows`
7218
7219 Whether the agent can start workflow runs.
7220
7221 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
7222
7223 The agent can start workflow runs.
7224
7225 - `type: "enabled"`
7226
7227 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
7228
7229 Whether a run's plan can define inline agents, which are not saved.
7230
7231 - `predefined_agents: array of BetaManagedAgentsAgentReference`
7232
7233 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
7234
7235 - `type: "agent"`
7236
7237 - `id: string`
7238
7239 - `version: number`
7240
7241 format: int32
7242
7243 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
7244
7245 The agent cannot start workflow runs.
7246
7247 - `type: "disabled"`
63287248
63297249 - `name: string`
63307250
from line 10252
933210252
933310253 - `"claude-haiku-4-5"`
933410254
9335 Fastest model with near-frontier intelligence
9336
933710255 - `"claude-haiku-4-5-20251001"`
933810256
9339 Fastest model with near-frontier intelligence
9340
934110257 - `"claude-opus-4-5"`
934210258
934310259 Powerful intelligence for long-running agents and coding
from line 10334
941810334
941910335 - `"claude-haiku-4-5"`
942010336
9421 Fastest model with near-frontier intelligence
9422
942310337 - `"claude-haiku-4-5-20251001"`
942410338
9425 Fastest model with near-frontier intelligence
9426
942710339 - `"claude-opus-4-5"`
942810340
942910341 Powerful intelligence for long-running agents and coding
from line 10462
955010462
955110463 - `"claude-haiku-4-5"`
955210464
9553 Fastest model with near-frontier intelligence
9554
955510465 - `"claude-haiku-4-5-20251001"`
955610466
9557 Fastest model with near-frontier intelligence
9558
955910467 - `"claude-opus-4-5"`
956010468
956110469 Powerful intelligence for long-running agents and coding
from line 10552
964410552
964510553 - `"fast"`
964610554
10555### Beta Managed Agents Multiagent Advisor
10556
10557- `BetaManagedAgentsMultiagentAdvisor = BetaManagedAgentsMultiagentAdvisorEnabled or BetaManagedAgentsMultiagentAdvisorDisabled`
10558
10559 Whether the session's primary thread can consult an advisor model.
10560
10561 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
10562
10563 The session's primary thread can consult `model` mid-turn.
10564
10565 - `type: "enabled"`
10566
10567 - `model: string`
10568
10569 The advisor model id.
10570
10571 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
10572
10573 The agent has no advisor.
10574
10575 - `type: "disabled"`
10576
10577### Beta Managed Agents Multiagent Advisor Disabled
10578
10579- `BetaManagedAgentsMultiagentAdvisorDisabled object`
10580
10581 The agent has no advisor.
10582
10583 - `type: "disabled"`
10584
10585### Beta Managed Agents Multiagent Advisor Disabled Params
10586
10587- `BetaManagedAgentsMultiagentAdvisorDisabledParams object`
10588
10589 The agent has no advisor.
10590
10591 - `type: "disabled"`
10592
10593### Beta Managed Agents Multiagent Advisor Enabled
10594
10595- `BetaManagedAgentsMultiagentAdvisorEnabled object`
10596
10597 The session's primary thread can consult `model` mid-turn.
10598
10599 - `type: "enabled"`
10600
10601 - `model: string`
10602
10603 The advisor model id.
10604
10605### Beta Managed Agents Multiagent Advisor Enabled Params
10606
10607- `BetaManagedAgentsMultiagentAdvisorEnabledParams object`
10608
10609 The session's primary thread can consult `model` mid-turn.
10610
10611 - `type: "enabled"`
10612
10613 - `model: string`
10614
10615 A Claude model id. The model must be permitted as an advisor for this agent's model.
10616
10617 minLength: 1, maxLength: 256
10618
10619### Beta Managed Agents Multiagent Advisor Params
10620
10621- `BetaManagedAgentsMultiagentAdvisorParams = BetaManagedAgentsMultiagentAdvisorEnabledParams or BetaManagedAgentsMultiagentAdvisorDisabledParams`
10622
10623 Whether the session's primary thread can consult an advisor model.
10624
10625 - `BetaManagedAgentsMultiagentAdvisorEnabledParams object`
10626
10627 The session's primary thread can consult `model` mid-turn.
10628
10629 - `type: "enabled"`
10630
10631 - `model: string`
10632
10633 A Claude model id. The model must be permitted as an advisor for this agent's model.
10634
10635 minLength: 1, maxLength: 256
10636
10637 - `BetaManagedAgentsMultiagentAdvisorDisabledParams object`
10638
10639 The agent has no advisor.
10640
10641 - `type: "disabled"`
10642
964710643### Beta Managed Agents Multiagent Coordinator
964810644
964910645- `BetaManagedAgentsMultiagentCoordinator object`
from line 10724
972810724
972910725 minLength: 1, maxLength: 256
973010726
10727### Beta Managed Agents Multiagent Inline Agents
10728
10729- `BetaManagedAgentsMultiagentInlineAgents = BetaManagedAgentsMultiagentInlineAgentsEnabled or BetaManagedAgentsMultiagentInlineAgentsDisabled`
10730
10731 Whether the agent can define inline agents. The agent defines an inline agent itself, in a workflow run's plan or when it spawns a session thread, and the inline agent is not saved.
10732
10733 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
10734
10735 The agent can define inline agents.
10736
10737 - `type: "enabled"`
10738
10739 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
10740
10741 The agent cannot define inline agents.
10742
10743 - `type: "disabled"`
10744
10745### Beta Managed Agents Multiagent Inline Agents Disabled
10746
10747- `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
10748
10749 The agent cannot define inline agents.
10750
10751 - `type: "disabled"`
10752
10753### Beta Managed Agents Multiagent Inline Agents Disabled Params
10754
10755- `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
10756
10757 The agent cannot define inline agents.
10758
10759 - `type: "disabled"`
10760
10761### Beta Managed Agents Multiagent Inline Agents Enabled
10762
10763- `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
10764
10765 The agent can define inline agents.
10766
10767 - `type: "enabled"`
10768
10769### Beta Managed Agents Multiagent Inline Agents Enabled Params
10770
10771- `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
10772
10773 The agent can define inline agents.
10774
10775 - `type: "enabled"`
10776
10777### Beta Managed Agents Multiagent Inline Agents Params
10778
10779- `BetaManagedAgentsMultiagentInlineAgentsParams = BetaManagedAgentsMultiagentInlineAgentsEnabledParams or BetaManagedAgentsMultiagentInlineAgentsDisabledParams`
10780
10781 Whether the agent can define inline agents. The agent defines an inline agent itself, in a workflow run's plan or when it spawns a session thread, and the inline agent is not saved.
10782
10783 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
10784
10785 The agent can define inline agents.
10786
10787 - `type: "enabled"`
10788
10789 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
10790
10791 The agent cannot define inline agents.
10792
10793 - `type: "disabled"`
10794
10795### Beta Managed Agents Multiagent Predefined Agent Params
10796
10797- `BetaManagedAgentsMultiagentPredefinedAgentParams = string or BetaManagedAgentsAgentParams or BetaManagedAgentsMultiagentSelfParams`
10798
10799 One agent in a `predefined_agents` list. It is an agent ID string, an `agent` reference with an optional `version`, or `self` for the agent that owns this configuration.
10800
10801 - `string`
10802
10803 - `BetaManagedAgentsAgentParams object`
10804
10805 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
10806
10807 - `type: "agent"`
10808
10809 - `id: string`
10810
10811 The `agent` ID.
10812
10813 minLength: 1, maxLength: 128
10814
10815 - `version: optional number`
10816
10817 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
10818
10819 format: int32
10820
10821 - `BetaManagedAgentsMultiagentSelfParams object`
10822
10823 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
10824
10825 - `type: "self"`
10826
973110827### Beta Managed Agents Multiagent Self Params
973210828
973310829- `BetaManagedAgentsMultiagentSelfParams object`
from line 10832
973610832
973710833 - `type: "self"`
973810834
10835### Beta Managed Agents Multiagent Subagents
10836
10837- `BetaManagedAgentsMultiagentSubagents = BetaManagedAgentsMultiagentSubagentsEnabled or BetaManagedAgentsMultiagentSubagentsDisabled`
10838
10839 Whether the agent can spawn session threads.
10840
10841 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
10842
10843 The agent can spawn session threads.
10844
10845 - `type: "enabled"`
10846
10847 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
10848
10849 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
10850
10851 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
10852
10853 The agent can define inline agents.
10854
10855 - `type: "enabled"`
10856
10857 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
10858
10859 The agent cannot define inline agents.
10860
10861 - `type: "disabled"`
10862
10863 - `predefined_agents: array of BetaManagedAgentsAgentReference`
10864
10865 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
10866
10867 - `type: "agent"`
10868
10869 - `id: string`
10870
10871 - `version: number`
10872
10873 format: int32
10874
10875 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
10876
10877 The agent cannot spawn session threads.
10878
10879 - `type: "disabled"`
10880
10881### Beta Managed Agents Multiagent Subagents Disabled
10882
10883- `BetaManagedAgentsMultiagentSubagentsDisabled object`
10884
10885 The agent cannot spawn session threads.
10886
10887 - `type: "disabled"`
10888
10889### Beta Managed Agents Multiagent Subagents Disabled Params
10890
10891- `BetaManagedAgentsMultiagentSubagentsDisabledParams object`
10892
10893 The agent cannot spawn session threads.
10894
10895 - `type: "disabled"`
10896
10897### Beta Managed Agents Multiagent Subagents Enabled
10898
10899- `BetaManagedAgentsMultiagentSubagentsEnabled object`
10900
10901 The agent can spawn session threads.
10902
10903 - `type: "enabled"`
10904
10905 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
10906
10907 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
10908
10909 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
10910
10911 The agent can define inline agents.
10912
10913 - `type: "enabled"`
10914
10915 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
10916
10917 The agent cannot define inline agents.
10918
10919 - `type: "disabled"`
10920
10921 - `predefined_agents: array of BetaManagedAgentsAgentReference`
10922
10923 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
10924
10925 - `type: "agent"`
10926
10927 - `id: string`
10928
10929 - `version: number`
10930
10931 format: int32
10932
10933### Beta Managed Agents Multiagent Subagents Enabled Params
10934
10935- `BetaManagedAgentsMultiagentSubagentsEnabledParams object`
10936
10937 The agent can spawn session threads. Each thread runs a predefined agent, which is a saved agent in `predefined_agents`, or an inline agent, which the agent defines when it spawns the thread and which is not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
10938
10939 - `type: "enabled"`
10940
10941 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
10942
10943 Whether the agent can define inline agents when it spawns session threads. Defaults to enabled.
10944
10945 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
10946
10947 The agent can define inline agents.
10948
10949 - `type: "enabled"`
10950
10951 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
10952
10953 The agent cannot define inline agents.
10954
10955 - `type: "disabled"`
10956
10957 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
10958
10959 Predefined agents that this agent can spawn as session threads. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `workflows.predefined_agents`, and an agent in one list is not added to the other.
10960
10961 - `string`
10962
10963 - `BetaManagedAgentsAgentParams object`
10964
10965 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
10966
10967 - `type: "agent"`
10968
10969 - `id: string`
10970
10971 The `agent` ID.
10972
10973 minLength: 1, maxLength: 128
10974
10975 - `version: optional number`
10976
10977 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
10978
10979 format: int32
10980
10981 - `BetaManagedAgentsMultiagentSelfParams object`
10982
10983 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
10984
10985 - `type: "self"`
10986
10987### Beta Managed Agents Multiagent Subagents Params
10988
10989- `BetaManagedAgentsMultiagentSubagentsParams = BetaManagedAgentsMultiagentSubagentsEnabledParams or BetaManagedAgentsMultiagentSubagentsDisabledParams`
10990
10991 Whether the agent can spawn session threads.
10992
10993 - `BetaManagedAgentsMultiagentSubagentsEnabledParams object`
10994
10995 The agent can spawn session threads. Each thread runs a predefined agent, which is a saved agent in `predefined_agents`, or an inline agent, which the agent defines when it spawns the thread and which is not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
10996
10997 - `type: "enabled"`
10998
10999 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
11000
11001 Whether the agent can define inline agents when it spawns session threads. Defaults to enabled.
11002
11003 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
11004
11005 The agent can define inline agents.
11006
11007 - `type: "enabled"`
11008
11009 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
11010
11011 The agent cannot define inline agents.
11012
11013 - `type: "disabled"`
11014
11015 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
11016
11017 Predefined agents that this agent can spawn as session threads. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `workflows.predefined_agents`, and an agent in one list is not added to the other.
11018
11019 - `string`
11020
11021 - `BetaManagedAgentsAgentParams object`
11022
11023 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
11024
11025 - `type: "agent"`
11026
11027 - `id: string`
11028
11029 The `agent` ID.
11030
11031 minLength: 1, maxLength: 128
11032
11033 - `version: optional number`
11034
11035 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
11036
11037 format: int32
11038
11039 - `BetaManagedAgentsMultiagentSelfParams object`
11040
11041 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
11042
11043 - `type: "self"`
11044
11045 - `BetaManagedAgentsMultiagentSubagentsDisabledParams object`
11046
11047 The agent cannot spawn session threads.
11048
11049 - `type: "disabled"`
11050
11051### Beta Managed Agents Multiagent Workflows
11052
11053- `BetaManagedAgentsMultiagentWorkflows = BetaManagedAgentsMultiagentWorkflowsEnabled or BetaManagedAgentsMultiagentWorkflowsDisabled`
11054
11055 Whether the agent can start workflow runs.
11056
11057 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
11058
11059 The agent can start workflow runs.
11060
11061 - `type: "enabled"`
11062
11063 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
11064
11065 Whether a run's plan can define inline agents, which are not saved.
11066
11067 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
11068
11069 The agent can define inline agents.
11070
11071 - `type: "enabled"`
11072
11073 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
11074
11075 The agent cannot define inline agents.
11076
11077 - `type: "disabled"`
11078
11079 - `predefined_agents: array of BetaManagedAgentsAgentReference`
11080
11081 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
11082
11083 - `type: "agent"`
11084
11085 - `id: string`
11086
11087 - `version: number`
11088
11089 format: int32
11090
11091 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
11092
11093 The agent cannot start workflow runs.
11094
11095 - `type: "disabled"`
11096
11097### Beta Managed Agents Multiagent Workflows Disabled
11098
11099- `BetaManagedAgentsMultiagentWorkflowsDisabled object`
11100
11101 The agent cannot start workflow runs.
11102
11103 - `type: "disabled"`
11104
11105### Beta Managed Agents Multiagent Workflows Disabled Params
11106
11107- `BetaManagedAgentsMultiagentWorkflowsDisabledParams object`
11108
11109 The agent cannot start workflow runs.
11110
11111 - `type: "disabled"`
11112
11113### Beta Managed Agents Multiagent Workflows Enabled
11114
11115- `BetaManagedAgentsMultiagentWorkflowsEnabled object`
11116
11117 The agent can start workflow runs.
11118
11119 - `type: "enabled"`
11120
11121 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
11122
11123 Whether a run's plan can define inline agents, which are not saved.
11124
11125 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
11126
11127 The agent can define inline agents.
11128
11129 - `type: "enabled"`
11130
11131 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
11132
11133 The agent cannot define inline agents.
11134
11135 - `type: "disabled"`
11136
11137 - `predefined_agents: array of BetaManagedAgentsAgentReference`
11138
11139 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
11140
11141 - `type: "agent"`
11142
11143 - `id: string`
11144
11145 - `version: number`
11146
11147 format: int32
11148
11149### Beta Managed Agents Multiagent Workflows Enabled Params
11150
11151- `BetaManagedAgentsMultiagentWorkflowsEnabledParams object`
11152
11153 The agent can start workflow runs. Each run follows a plan, a program that the agent writes. A plan can use predefined agents, which are the saved agents in `predefined_agents`, and inline agents, which it defines itself and which are not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
11154
11155 - `type: "enabled"`
11156
11157 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
11158
11159 Whether a run's plan can define inline agents. Defaults to enabled.
11160
11161 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
11162
11163 The agent can define inline agents.
11164
11165 - `type: "enabled"`
11166
11167 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
11168
11169 The agent cannot define inline agents.
11170
11171 - `type: "disabled"`
11172
11173 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
11174
11175 Predefined agents that a run's plan can use. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `subagents.predefined_agents`, and an agent in one list is not added to the other.
11176
11177 - `string`
11178
11179 - `BetaManagedAgentsAgentParams object`
11180
11181 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
11182
11183 - `type: "agent"`
11184
11185 - `id: string`
11186
11187 The `agent` ID.
11188
11189 minLength: 1, maxLength: 128
11190
11191 - `version: optional number`
11192
11193 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
11194
11195 format: int32
11196
11197 - `BetaManagedAgentsMultiagentSelfParams object`
11198
11199 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
11200
11201 - `type: "self"`
11202
11203### Beta Managed Agents Multiagent Workflows Params
11204
11205- `BetaManagedAgentsMultiagentWorkflowsParams = BetaManagedAgentsMultiagentWorkflowsEnabledParams or BetaManagedAgentsMultiagentWorkflowsDisabledParams`
11206
11207 Whether the agent can start workflow runs.
11208
11209 - `BetaManagedAgentsMultiagentWorkflowsEnabledParams object`
11210
11211 The agent can start workflow runs. Each run follows a plan, a program that the agent writes. A plan can use predefined agents, which are the saved agents in `predefined_agents`, and inline agents, which it defines itself and which are not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
11212
11213 - `type: "enabled"`
11214
11215 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
11216
11217 Whether a run's plan can define inline agents. Defaults to enabled.
11218
11219 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
11220
11221 The agent can define inline agents.
11222
11223 - `type: "enabled"`
11224
11225 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
11226
11227 The agent cannot define inline agents.
11228
11229 - `type: "disabled"`
11230
11231 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
11232
11233 Predefined agents that a run's plan can use. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `subagents.predefined_agents`, and an agent in one list is not added to the other.
11234
11235 - `string`
11236
11237 - `BetaManagedAgentsAgentParams object`
11238
11239 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
11240
11241 - `type: "agent"`
11242
11243 - `id: string`
11244
11245 The `agent` ID.
11246
11247 minLength: 1, maxLength: 128
11248
11249 - `version: optional number`
11250
11251 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
11252
11253 format: int32
11254
11255 - `BetaManagedAgentsMultiagentSelfParams object`
11256
11257 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
11258
11259 - `type: "self"`
11260
11261 - `BetaManagedAgentsMultiagentWorkflowsDisabledParams object`
11262
11263 The agent cannot start workflow runs.
11264
11265 - `type: "disabled"`
11266
11267### Beta Managed Agents Multiagent20261001
11268
11269- `BetaManagedAgentsMultiagent20261001 object`
11270
11271 Resolved multiagent configuration with three members, each enabled or disabled on its own.
11272
11273 - `type: "multiagent_20261001"`
11274
11275 - `advisor: BetaManagedAgentsMultiagentAdvisor`
11276
11277 Whether the session's primary thread can consult an advisor model.
11278
11279 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
11280
11281 The session's primary thread can consult `model` mid-turn.
11282
11283 - `type: "enabled"`
11284
11285 - `model: string`
11286
11287 The advisor model id.
11288
11289 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
11290
11291 The agent has no advisor.
11292
11293 - `type: "disabled"`
11294
11295 - `subagents: BetaManagedAgentsMultiagentSubagents`
11296
11297 Whether the agent can spawn session threads.
11298
11299 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
11300
11301 The agent can spawn session threads.
11302
11303 - `type: "enabled"`
11304
11305 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
11306
11307 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
11308
11309 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
11310
11311 The agent can define inline agents.
11312
11313 - `type: "enabled"`
11314
11315 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
11316
11317 The agent cannot define inline agents.
11318
11319 - `type: "disabled"`
11320
11321 - `predefined_agents: array of BetaManagedAgentsAgentReference`
11322
11323 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
11324
11325 - `type: "agent"`
11326
11327 - `id: string`
11328
11329 - `version: number`
11330
11331 format: int32
11332
11333 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
11334
11335 The agent cannot spawn session threads.
11336
11337 - `type: "disabled"`
11338
11339 - `workflows: BetaManagedAgentsMultiagentWorkflows`
11340
11341 Whether the agent can start workflow runs.
11342
11343 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
11344
11345 The agent can start workflow runs.
11346
11347 - `type: "enabled"`
11348
11349 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
11350
11351 Whether a run's plan can define inline agents, which are not saved.
11352
11353 - `predefined_agents: array of BetaManagedAgentsAgentReference`
11354
11355 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
11356
11357 - `type: "agent"`
11358
11359 - `id: string`
11360
11361 - `version: number`
11362
11363 format: int32
11364
11365 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
11366
11367 The agent cannot start workflow runs.
11368
11369 - `type: "disabled"`
11370
11371### Beta Managed Agents Multiagent20261001 Params
11372
11373- `BetaManagedAgentsMultiagent20261001Params object`
11374
11375 Multiagent configuration with three members, each enabled or disabled on its own. On an update, if the agent's stored `multiagent` also has type `multiagent_20261001`, this configuration is merged into the stored one, level by level, instead of replacing it. A key that the update omits keeps its stored value. A key sent as null takes its default, on create as well, so `"workflows": null` enables workflows. An object sent with a `type` other than the stored one replaces the stored object, and the keys that it omits take their defaults. A `predefined_agents` list that is sent replaces the stored list. Every object that is sent needs its `type`, and an enabled `advisor` needs its `model`. Other validation applies to the merged result.
11376
11377 - `type: "multiagent_20261001"`
11378
11379 - `advisor: optional BetaManagedAgentsMultiagentAdvisorParams or null`
11380
11381 Whether the session's primary thread can consult an advisor model. Defaults to disabled.
11382
11383 - `BetaManagedAgentsMultiagentAdvisorEnabledParams object`
11384
11385 The session's primary thread can consult `model` mid-turn.
11386
11387 - `type: "enabled"`
11388
11389 - `model: string`
11390
11391 A Claude model id. The model must be permitted as an advisor for this agent's model.
11392
11393 minLength: 1, maxLength: 256
11394
11395 - `BetaManagedAgentsMultiagentAdvisorDisabledParams object`
11396
11397 The agent has no advisor.
11398
11399 - `type: "disabled"`
11400
11401 - `subagents: optional BetaManagedAgentsMultiagentSubagentsParams or null`
11402
11403 Whether the agent can spawn session threads. Defaults to enabled.
11404
11405 - `BetaManagedAgentsMultiagentSubagentsEnabledParams object`
11406
11407 The agent can spawn session threads. Each thread runs a predefined agent, which is a saved agent in `predefined_agents`, or an inline agent, which the agent defines when it spawns the thread and which is not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
11408
11409 - `type: "enabled"`
11410
11411 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
11412
11413 Whether the agent can define inline agents when it spawns session threads. Defaults to enabled.
11414
11415 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
11416
11417 The agent can define inline agents.
11418
11419 - `type: "enabled"`
11420
11421 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
11422
11423 The agent cannot define inline agents.
11424
11425 - `type: "disabled"`
11426
11427 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
11428
11429 Predefined agents that this agent can spawn as session threads. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `workflows.predefined_agents`, and an agent in one list is not added to the other.
11430
11431 - `string`
11432
11433 - `BetaManagedAgentsAgentParams object`
11434
11435 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
11436
11437 - `type: "agent"`
11438
11439 - `id: string`
11440
11441 The `agent` ID.
11442
11443 minLength: 1, maxLength: 128
11444
11445 - `version: optional number`
11446
11447 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
11448
11449 format: int32
11450
11451 - `BetaManagedAgentsMultiagentSelfParams object`
11452
11453 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
11454
11455 - `type: "self"`
11456
11457 - `BetaManagedAgentsMultiagentSubagentsDisabledParams object`
11458
11459 The agent cannot spawn session threads.
11460
11461 - `type: "disabled"`
11462
11463 - `workflows: optional BetaManagedAgentsMultiagentWorkflowsParams or null`
11464
11465 Whether the agent can start workflow runs. Defaults to enabled.
11466
11467 - `BetaManagedAgentsMultiagentWorkflowsEnabledParams object`
11468
11469 The agent can start workflow runs. Each run follows a plan, a program that the agent writes. A plan can use predefined agents, which are the saved agents in `predefined_agents`, and inline agents, which it defines itself and which are not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
11470
11471 - `type: "enabled"`
11472
11473 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
11474
11475 Whether a run's plan can define inline agents. Defaults to enabled.
11476
11477 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
11478
11479 Predefined agents that a run's plan can use. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `subagents.predefined_agents`, and an agent in one list is not added to the other.
11480
11481 - `string`
11482
11483 - `BetaManagedAgentsAgentParams object`
11484
11485 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
11486
11487 - `BetaManagedAgentsMultiagentSelfParams object`
11488
11489 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
11490
11491 - `BetaManagedAgentsMultiagentWorkflowsDisabledParams object`
11492
11493 The agent cannot start workflow runs.
11494
11495 - `type: "disabled"`
11496
973911497### Beta Managed Agents Read Tool Config
974011498
974111499- `BetaManagedAgentsReadToolConfig object`
from line 11642
988411642
988511643 - `"claude-haiku-4-5"`
988611644
9887 Fastest model with near-frontier intelligence
9888
988911645 - `"claude-haiku-4-5-20251001"`
989011646
9891 Fastest model with near-frontier intelligence
9892
989311647 - `"claude-opus-4-5"`
989411648
989511649 Powerful intelligence for long-running agents and coding
from line 13420
1166613420
1166713421 - `"claude-haiku-4-5"`
1166813422
11669 Fastest model with near-frontier intelligence
11670
1167113423 - `"claude-haiku-4-5-20251001"`
1167213424
11673 Fastest model with near-frontier intelligence
11674
1167513425 - `"claude-opus-4-5"`
1167613426
1167713427 Powerful intelligence for long-running agents and coding
from line 13494
1174413494
1174513495 Multiagent orchestration configuration. Null when the agent is single-threaded.
1174613496
11747 - `type: "coordinator"`
11748
11749 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
11750
11751 Agents the coordinator may spawn as session threads, each resolved to a specific version.
11752
11753 - `BetaManagedAgentsAgentReference object`
11754
11755 A resolved agent reference with a concrete version.
11756
11757 - `type: "agent"`
11758
11759 - `id: string`
11760
11761 - `version: number`
11762
11763 format: int32
11764
11765 - `BetaManagedAgentsAdvisor object`
11766
11767 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
11768
11769 - `type: "advisor"`
11770
11771 - `model: string`
11772
11773 The advisor model id.
13497 - `BetaManagedAgentsMultiagentCoordinator object`
13498
13499 Resolved coordinator topology with a concrete agent roster.
13500
13501 - `type: "coordinator"`
13502
13503 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
13504
13505 Agents the coordinator may spawn as session threads, each resolved to a specific version.
13506
13507 - `BetaManagedAgentsAgentReference object`
13508
13509 A resolved agent reference with a concrete version.
13510
13511 - `type: "agent"`
13512
13513 - `id: string`
13514
13515 - `version: number`
13516
13517 format: int32
13518
13519 - `BetaManagedAgentsAdvisor object`
13520
13521 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
13522
13523 - `type: "advisor"`
13524
13525 - `model: string`
13526
13527 The advisor model id.
13528
13529 - `BetaManagedAgentsMultiagent20261001 object`
13530
13531 Resolved multiagent configuration with three members, each enabled or disabled on its own.
13532
13533 - `type: "multiagent_20261001"`
13534
13535 - `advisor: BetaManagedAgentsMultiagentAdvisor`
13536
13537 Whether the session's primary thread can consult an advisor model.
13538
13539 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
13540
13541 The session's primary thread can consult `model` mid-turn.
13542
13543 - `type: "enabled"`
13544
13545 - `model: string`
13546
13547 The advisor model id.
13548
13549 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
13550
13551 The agent has no advisor.
13552
13553 - `type: "disabled"`
13554
13555 - `subagents: BetaManagedAgentsMultiagentSubagents`
13556
13557 Whether the agent can spawn session threads.
13558
13559 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
13560
13561 The agent can spawn session threads.
13562
13563 - `type: "enabled"`
13564
13565 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
13566
13567 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
13568
13569 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
13570
13571 The agent can define inline agents.
13572
13573 - `type: "enabled"`
13574
13575 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
13576
13577 The agent cannot define inline agents.
13578
13579 - `type: "disabled"`
13580
13581 - `predefined_agents: array of BetaManagedAgentsAgentReference`
13582
13583 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
13584
13585 - `type: "agent"`
13586
13587 - `id: string`
13588
13589 - `version: number`
13590
13591 format: int32
13592
13593 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
13594
13595 The agent cannot spawn session threads.
13596
13597 - `type: "disabled"`
13598
13599 - `workflows: BetaManagedAgentsMultiagentWorkflows`
13600
13601 Whether the agent can start workflow runs.
13602
13603 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
13604
13605 The agent can start workflow runs.
13606
13607 - `type: "enabled"`
13608
13609 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
13610
13611 Whether a run's plan can define inline agents, which are not saved.
13612
13613 - `predefined_agents: array of BetaManagedAgentsAgentReference`
13614
13615 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
13616
13617 - `type: "agent"`
13618
13619 - `id: string`
13620
13621 - `version: number`
13622
13623 format: int32
13624
13625 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
13626
13627 The agent cannot start workflow runs.
13628
13629 - `type: "disabled"`
1177413630
1177513631 - `name: string`
1177613632
from line 14140
1228414140 "speed": "standard"
1228514141 },
1228614142 "multiagent": {
12287 "agents": [
12288 {
12289 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
12290 "type": "agent",
12291 "version": 1
12292 }
12293 ],
12294 "type": "coordinator"
14143 "advisor": {
14144 "type": "disabled"
14145 },
14146 "subagents": {
14147 "inline_agents": {
14148 "type": "enabled"
14149 },
14150 "predefined_agents": [
14151 {
14152 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
14153 "type": "agent",
14154 "version": 1
14155 }
14156 ],
14157 "type": "enabled"
14158 },
14159 "type": "multiagent_20261001",
14160 "workflows": {
14161 "inline_agents": {
14162 "type": "enabled"
14163 },
14164 "predefined_agents": [
14165 {
14166 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
14167 "type": "agent",
14168 "version": 1
14169 }
14170 ],
14171 "type": "enabled"
14172 }
1229514173 },
1229614174 "name": "My First Agent",
1229714175 "skills": [
1229814176
api/beta/agents/archive Changed · +163 / -39 lines
from line 221
221221
222222 - `"claude-haiku-4-5"`
223223
224 Fastest model with near-frontier intelligence
225
226224 - `"claude-haiku-4-5-20251001"`
227225
228 Fastest model with near-frontier intelligence
229
230226 - `"claude-opus-4-5"`
231227
232228 Powerful intelligence for long-running agents and coding
from line 295
299295
300296 Multiagent orchestration configuration. Null when the agent is single-threaded.
301297
302 - `type: "coordinator"`
298 - `BetaManagedAgentsMultiagentCoordinator object`
303299
304 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
300 Resolved coordinator topology with a concrete agent roster.
305301
306 Agents the coordinator may spawn as session threads, each resolved to a specific version.
302 - `type: "coordinator"`
307303
308 - `BetaManagedAgentsAgentReference object`
304 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
309305
310 A resolved agent reference with a concrete version.
306 Agents the coordinator may spawn as session threads, each resolved to a specific version.
311307
312 - `type: "agent"`
308 - `BetaManagedAgentsAgentReference object`
313309
314 - `id: string`
310 A resolved agent reference with a concrete version.
315311
316 - `version: number`
312 - `type: "agent"`
317313
318 format: int32
314 - `id: string`
319315
320 - `BetaManagedAgentsAdvisor object`
316 - `version: number`
321317
322 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
318 format: int32
323319
324 - `type: "advisor"`
320 - `BetaManagedAgentsAdvisor object`
325321
326 - `model: string`
322 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
327323
328 The advisor model id.
324 - `type: "advisor"`
329325
326 - `model: string`
327
328 The advisor model id.
329
330 - `BetaManagedAgentsMultiagent20261001 object`
331
332 Resolved multiagent configuration with three members, each enabled or disabled on its own.
333
334 - `type: "multiagent_20261001"`
335
336 - `advisor: BetaManagedAgentsMultiagentAdvisor`
337
338 Whether the session's primary thread can consult an advisor model.
339
340 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
341
342 The session's primary thread can consult `model` mid-turn.
343
344 - `type: "enabled"`
345
346 - `model: string`
347
348 The advisor model id.
349
350 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
351
352 The agent has no advisor.
353
354 - `type: "disabled"`
355
356 - `subagents: BetaManagedAgentsMultiagentSubagents`
357
358 Whether the agent can spawn session threads.
359
360 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
361
362 The agent can spawn session threads.
363
364 - `type: "enabled"`
365
366 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
367
368 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
369
370 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
371
372 The agent can define inline agents.
373
374 - `type: "enabled"`
375
376 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
377
378 The agent cannot define inline agents.
379
380 - `type: "disabled"`
381
382 - `predefined_agents: array of BetaManagedAgentsAgentReference`
383
384 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
385
386 - `type: "agent"`
387
388 - `id: string`
389
390 - `version: number`
391
392 format: int32
393
394 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
395
396 The agent cannot spawn session threads.
397
398 - `type: "disabled"`
399
400 - `workflows: BetaManagedAgentsMultiagentWorkflows`
401
402 Whether the agent can start workflow runs.
403
404 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
405
406 The agent can start workflow runs.
407
408 - `type: "enabled"`
409
410 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
411
412 Whether a run's plan can define inline agents, which are not saved.
413
414 - `predefined_agents: array of BetaManagedAgentsAgentReference`
415
416 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
417
418 - `type: "agent"`
419
420 - `id: string`
421
422 - `version: number`
423
424 format: int32
425
426 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
427
428 The agent cannot start workflow runs.
429
430 - `type: "disabled"`
431
330432 - `name: string`
331433
332434 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 936
834936 "speed": "standard"
835937 },
836938 "multiagent": {
837 "agents": [
838 {
839 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
840 "type": "agent",
841 "version": 1
842 }
843 ],
844 "type": "coordinator"
939 "advisor": {
940 "type": "disabled"
941 },
942 "subagents": {
943 "inline_agents": {
944 "type": "enabled"
945 },
946 "predefined_agents": [
947 {
948 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
949 "type": "agent",
950 "version": 1
951 }
952 ],
953 "type": "enabled"
954 },
955 "type": "multiagent_20261001",
956 "workflows": {
957 "inline_agents": {
958 "type": "enabled"
959 },
960 "predefined_agents": [
961 {
962 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
963 "type": "agent",
964 "version": 1
965 }
966 ],
967 "type": "enabled"
968 }
845969 },
846970 "name": "My First Agent",
847971 "skills": [
api/beta/agents/create Changed · +863 / -640 lines
from line 183
183183
184184 - `"claude-haiku-4-5"`
185185
186 Fastest model with near-frontier intelligence
187
188186 - `"claude-haiku-4-5-20251001"`
189187
190 Fastest model with near-frontier intelligence
191
192188 - `"claude-opus-4-5"`
193189
194190 Powerful intelligence for long-running agents and coding
from line 321
325321
326322 Multiagent orchestration configuration.
327323
328 - `type: "coordinator"`
324 - `BetaManagedAgentsMultiagentCoordinatorParams object`
329325
330 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
326 A coordinator topology: the session's primary thread orchestrates work by spawning session threads, each running an agent drawn from the `agents` roster.
331327
332 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
328 - `type: "coordinator"`
333329
334 - `string`
330 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
335331
336 - `BetaManagedAgentsAgentParams object`
332 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
337333
338 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
334 - `string`
339335
340 - `type: "agent"`
336 - `BetaManagedAgentsAgentParams object`
341337
342 - `id: string`
338 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
343339
344 The `agent` ID.
340 - `type: "agent"`
345341
346 minLength: 1, maxLength: 128
342 - `id: string`
347343
348 - `version: optional number`
344 The `agent` ID.
349345
350 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
346 minLength: 1, maxLength: 128
351347
352 format: int32
348 - `version: optional number`
353349
354 - `BetaManagedAgentsMultiagentSelfParams object`
350 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
355351
356 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
352 format: int32
357353
358 - `type: "self"`
354 - `BetaManagedAgentsMultiagentSelfParams object`
359355
360 - `BetaManagedAgentsAdvisorParams object`
356 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
361357
362 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
358 - `type: "self"`
363359
364 - `type: "advisor"`
360 - `BetaManagedAgentsAdvisorParams object`
365361
366 - `model: string`
362 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
367363
368 A Claude model id. The model must be permitted as an advisor for this agent's model.
364 - `type: "advisor"`
369365
370 minLength: 1, maxLength: 256
366 - `model: string`
371367
368 A Claude model id. The model must be permitted as an advisor for this agent's model.
369
370 minLength: 1, maxLength: 256
371
372 - `BetaManagedAgentsMultiagent20261001Params object`
373
374 Multiagent configuration with three members, each enabled or disabled on its own. On an update, if the agent's stored `multiagent` also has type `multiagent_20261001`, this configuration is merged into the stored one, level by level, instead of replacing it. A key that the update omits keeps its stored value. A key sent as null takes its default, on create as well, so `"workflows": null` enables workflows. An object sent with a `type` other than the stored one replaces the stored object, and the keys that it omits take their defaults. A `predefined_agents` list that is sent replaces the stored list. Every object that is sent needs its `type`, and an enabled `advisor` needs its `model`. Other validation applies to the merged result.
375
376 - `type: "multiagent_20261001"`
377
378 - `advisor: optional BetaManagedAgentsMultiagentAdvisorParams or null`
379
380 Whether the session's primary thread can consult an advisor model. Defaults to disabled.
381
382 - `BetaManagedAgentsMultiagentAdvisorEnabledParams object`
383
384 The session's primary thread can consult `model` mid-turn.
385
386 - `type: "enabled"`
387
388 - `model: string`
389
390 A Claude model id. The model must be permitted as an advisor for this agent's model.
391
392 minLength: 1, maxLength: 256
393
394 - `BetaManagedAgentsMultiagentAdvisorDisabledParams object`
395
396 The agent has no advisor.
397
398 - `type: "disabled"`
399
400 - `subagents: optional BetaManagedAgentsMultiagentSubagentsParams or null`
401
402 Whether the agent can spawn session threads. Defaults to enabled.
403
404 - `BetaManagedAgentsMultiagentSubagentsEnabledParams object`
405
406 The agent can spawn session threads. Each thread runs a predefined agent, which is a saved agent in `predefined_agents`, or an inline agent, which the agent defines when it spawns the thread and which is not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
407
408 - `type: "enabled"`
409
410 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
411
412 Whether the agent can define inline agents when it spawns session threads. Defaults to enabled.
413
414 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
415
416 The agent can define inline agents.
417
418 - `type: "enabled"`
419
420 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
421
422 The agent cannot define inline agents.
423
424 - `type: "disabled"`
425
426 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
427
428 Predefined agents that this agent can spawn as session threads. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `workflows.predefined_agents`, and an agent in one list is not added to the other.
429
430 - `string`
431
432 - `BetaManagedAgentsAgentParams object`
433
434 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
435
436 - `BetaManagedAgentsMultiagentSelfParams object`
437
438 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
439
440 - `BetaManagedAgentsMultiagentSubagentsDisabledParams object`
441
442 The agent cannot spawn session threads.
443
444 - `type: "disabled"`
445
446 - `workflows: optional BetaManagedAgentsMultiagentWorkflowsParams or null`
447
448 Whether the agent can start workflow runs. Defaults to enabled.
449
450 - `BetaManagedAgentsMultiagentWorkflowsEnabledParams object`
451
452 The agent can start workflow runs. Each run follows a plan, a program that the agent writes. A plan can use predefined agents, which are the saved agents in `predefined_agents`, and inline agents, which it defines itself and which are not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
453
454 - `type: "enabled"`
455
456 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
457
458 Whether a run's plan can define inline agents. Defaults to enabled.
459
460 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
461
462 Predefined agents that a run's plan can use. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `subagents.predefined_agents`, and an agent in one list is not added to the other.
463
464 - `string`
465
466 - `BetaManagedAgentsAgentParams object`
467
468 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
469
470 - `BetaManagedAgentsMultiagentSelfParams object`
471
472 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
473
474 - `BetaManagedAgentsMultiagentWorkflowsDisabledParams object`
475
476 The agent cannot start workflow runs.
477
478 - `type: "disabled"`
479
372480- `skills: optional array of BetaManagedAgentsSkillParams`
373481
374482 Skills available to the agent.
from line 1139
10311139
10321140 - `"claude-haiku-4-5"`
10331141
1034 Fastest model with near-frontier intelligence
1035
10361142 - `"claude-haiku-4-5-20251001"`
10371143
1038 Fastest model with near-frontier intelligence
1039
10401144 - `"claude-opus-4-5"`
10411145
10421146 Powerful intelligence for long-running agents and coding
from line 1213
11091213
11101214 Multiagent orchestration configuration. Null when the agent is single-threaded.
11111215
1112 - `type: "coordinator"`
1216 - `BetaManagedAgentsMultiagentCoordinator object`
11131217
1114 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
1218 Resolved coordinator topology with a concrete agent roster.
11151219
1116 Agents the coordinator may spawn as session threads, each resolved to a specific version.
1220 - `type: "coordinator"`
11171221
1118 - `BetaManagedAgentsAgentReference object`
1222 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
11191223
1120 A resolved agent reference with a concrete version.
1224 Agents the coordinator may spawn as session threads, each resolved to a specific version.
11211225
1122 - `type: "agent"`
1226 - `BetaManagedAgentsAgentReference object`
11231227
1124 - `id: string`
1228 A resolved agent reference with a concrete version.
11251229
1126 - `version: number`
1230 - `type: "agent"`
11271231
1128 format: int32
1232 - `id: string`
11291233
1130 - `BetaManagedAgentsAdvisor object`
1234 - `version: number`
11311235
1132 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
1236 format: int32
11331237
1134 - `type: "advisor"`
1238 - `BetaManagedAgentsAdvisor object`
11351239
1136 - `model: string`
1240 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
11371241
1138 The advisor model id.
1242 - `type: "advisor"`
11391243
1244 - `model: string`
1245
1246 The advisor model id.
1247
1248 - `BetaManagedAgentsMultiagent20261001 object`
1249
1250 Resolved multiagent configuration with three members, each enabled or disabled on its own.
1251
1252 - `type: "multiagent_20261001"`
1253
1254 - `advisor: BetaManagedAgentsMultiagentAdvisor`
1255
1256 Whether the session's primary thread can consult an advisor model.
1257
1258 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
1259
1260 The session's primary thread can consult `model` mid-turn.
1261
1262 - `type: "enabled"`
1263
1264 - `model: string`
1265
1266 The advisor model id.
1267
1268 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
1269
1270 The agent has no advisor.
1271
1272 - `type: "disabled"`
1273
1274 - `subagents: BetaManagedAgentsMultiagentSubagents`
1275
1276 Whether the agent can spawn session threads.
1277
1278 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
1279
1280 The agent can spawn session threads.
1281
1282 - `type: "enabled"`
1283
1284 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
1285
1286 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
1287
1288 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
1289
1290 The agent can define inline agents.
1291
1292 - `type: "enabled"`
1293
1294 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
1295
1296 The agent cannot define inline agents.
1297
1298 - `type: "disabled"`
1299
1300 - `predefined_agents: array of BetaManagedAgentsAgentReference`
1301
1302 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
1303
1304 - `type: "agent"`
1305
1306 - `id: string`
1307
1308 - `version: number`
1309
1310 format: int32
1311
1312 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
1313
1314 The agent cannot spawn session threads.
1315
1316 - `type: "disabled"`
1317
1318 - `workflows: BetaManagedAgentsMultiagentWorkflows`
1319
1320 Whether the agent can start workflow runs.
1321
1322 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
1323
1324 The agent can start workflow runs.
1325
1326 - `type: "enabled"`
1327
1328 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
1329
1330 Whether a run's plan can define inline agents, which are not saved.
1331
1332 - `predefined_agents: array of BetaManagedAgentsAgentReference`
1333
1334 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
1335
1336 - `type: "agent"`
1337
1338 - `id: string`
1339
1340 - `version: number`
1341
1342 format: int32
1343
1344 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
1345
1346 The agent cannot start workflow runs.
1347
1348 - `type: "disabled"`
1349
11401350 - `name: string`
11411351
11421352 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 1832
16221832 "metadata": {
16231833 "foo": "bar"
16241834 },
1625 "multiagent": {
1626 "agents": [
1627 "agent_011CZkYqphY8vELVzwCUpqiQ",
1628 {
1629 "type": "self"
1630 }
1631 ],
1632 "type": "coordinator"
1633 },
16341835 "system": "You are a general-purpose agent that can research, write code, run commands, and use connected tools to complete the user'\''s task end to end.",
16351836 "tools": [
16361837 {
from line 1868
16671868 "speed": "standard"
16681869 },
16691870 "multiagent": {
1670 "agents": [
1671 {
1672 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1673 "type": "agent",
1674 "version": 1
1675 }
1676 ],
1677 "type": "coordinator"
1871 "advisor": {
1872 "type": "disabled"
1873 },
1874 "subagents": {
1875 "inline_agents": {
1876 "type": "enabled"
1877 },
1878 "predefined_agents": [
1879 {
1880 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1881 "type": "agent",
1882 "version": 1
1883 }
1884 ],
1885 "type": "enabled"
1886 },
1887 "type": "multiagent_20261001",
1888 "workflows": {
1889 "inline_agents": {
1890 "type": "enabled"
1891 },
1892 "predefined_agents": [
1893 {
1894 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1895 "type": "agent",
1896 "version": 1
1897 }
1898 ],
1899 "type": "enabled"
1900 }
16781901 },
16791902 "name": "My First Agent",
16801903 "skills": [
api/beta/agents/list Changed · +163 / -39 lines
from line 243
243243
244244 - `"claude-haiku-4-5"`
245245
246 Fastest model with near-frontier intelligence
247
248246 - `"claude-haiku-4-5-20251001"`
249247
250 Fastest model with near-frontier intelligence
251
252248 - `"claude-opus-4-5"`
253249
254250 Powerful intelligence for long-running agents and coding
from line 317
321317
322318 Multiagent orchestration configuration. Null when the agent is single-threaded.
323319
324 - `type: "coordinator"`
320 - `BetaManagedAgentsMultiagentCoordinator object`
325321
326 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
322 Resolved coordinator topology with a concrete agent roster.
327323
328 Agents the coordinator may spawn as session threads, each resolved to a specific version.
324 - `type: "coordinator"`
329325
330 - `BetaManagedAgentsAgentReference object`
326 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
331327
332 A resolved agent reference with a concrete version.
328 Agents the coordinator may spawn as session threads, each resolved to a specific version.
333329
334 - `type: "agent"`
330 - `BetaManagedAgentsAgentReference object`
335331
336 - `id: string`
332 A resolved agent reference with a concrete version.
337333
338 - `version: number`
334 - `type: "agent"`
339335
340 format: int32
336 - `id: string`
341337
342 - `BetaManagedAgentsAdvisor object`
338 - `version: number`
343339
344 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
340 format: int32
345341
346 - `type: "advisor"`
342 - `BetaManagedAgentsAdvisor object`
347343
348 - `model: string`
344 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
349345
350 The advisor model id.
346 - `type: "advisor"`
351347
348 - `model: string`
349
350 The advisor model id.
351
352 - `BetaManagedAgentsMultiagent20261001 object`
353
354 Resolved multiagent configuration with three members, each enabled or disabled on its own.
355
356 - `type: "multiagent_20261001"`
357
358 - `advisor: BetaManagedAgentsMultiagentAdvisor`
359
360 Whether the session's primary thread can consult an advisor model.
361
362 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
363
364 The session's primary thread can consult `model` mid-turn.
365
366 - `type: "enabled"`
367
368 - `model: string`
369
370 The advisor model id.
371
372 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
373
374 The agent has no advisor.
375
376 - `type: "disabled"`
377
378 - `subagents: BetaManagedAgentsMultiagentSubagents`
379
380 Whether the agent can spawn session threads.
381
382 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
383
384 The agent can spawn session threads.
385
386 - `type: "enabled"`
387
388 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
389
390 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
391
392 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
393
394 The agent can define inline agents.
395
396 - `type: "enabled"`
397
398 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
399
400 The agent cannot define inline agents.
401
402 - `type: "disabled"`
403
404 - `predefined_agents: array of BetaManagedAgentsAgentReference`
405
406 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
407
408 - `type: "agent"`
409
410 - `id: string`
411
412 - `version: number`
413
414 format: int32
415
416 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
417
418 The agent cannot spawn session threads.
419
420 - `type: "disabled"`
421
422 - `workflows: BetaManagedAgentsMultiagentWorkflows`
423
424 Whether the agent can start workflow runs.
425
426 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
427
428 The agent can start workflow runs.
429
430 - `type: "enabled"`
431
432 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
433
434 Whether a run's plan can define inline agents, which are not saved.
435
436 - `predefined_agents: array of BetaManagedAgentsAgentReference`
437
438 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
439
440 - `type: "agent"`
441
442 - `id: string`
443
444 - `version: number`
445
446 format: int32
447
448 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
449
450 The agent cannot start workflow runs.
451
452 - `type: "disabled"`
453
352454 - `name: string`
353455
354456 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 963
861963 "speed": "standard"
862964 },
863965 "multiagent": {
864 "agents": [
865 {
866 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
867 "type": "agent",
868 "version": 1
869 }
870 ],
871 "type": "coordinator"
966 "advisor": {
967 "type": "disabled"
968 },
969 "subagents": {
970 "inline_agents": {
971 "type": "enabled"
972 },
973 "predefined_agents": [
974 {
975 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
976 "type": "agent",
977 "version": 1
978 }
979 ],
980 "type": "enabled"
981 },
982 "type": "multiagent_20261001",
983 "workflows": {
984 "inline_agents": {
985 "type": "enabled"
986 },
987 "predefined_agents": [
988 {
989 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
990 "type": "agent",
991 "version": 1
992 }
993 ],
994 "type": "enabled"
995 }
872996 },
873997 "name": "My First Agent",
874998 "skills": [
api/beta/agents/retrieve Changed · +163 / -39 lines
from line 229
229229
230230 - `"claude-haiku-4-5"`
231231
232 Fastest model with near-frontier intelligence
233
234232 - `"claude-haiku-4-5-20251001"`
235233
236 Fastest model with near-frontier intelligence
237
238234 - `"claude-opus-4-5"`
239235
240236 Powerful intelligence for long-running agents and coding
from line 303
307303
308304 Multiagent orchestration configuration. Null when the agent is single-threaded.
309305
310 - `type: "coordinator"`
306 - `BetaManagedAgentsMultiagentCoordinator object`
311307
312 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
308 Resolved coordinator topology with a concrete agent roster.
313309
314 Agents the coordinator may spawn as session threads, each resolved to a specific version.
310 - `type: "coordinator"`
315311
316 - `BetaManagedAgentsAgentReference object`
312 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
317313
318 A resolved agent reference with a concrete version.
314 Agents the coordinator may spawn as session threads, each resolved to a specific version.
319315
320 - `type: "agent"`
316 - `BetaManagedAgentsAgentReference object`
321317
322 - `id: string`
318 A resolved agent reference with a concrete version.
323319
324 - `version: number`
320 - `type: "agent"`
325321
326 format: int32
322 - `id: string`
327323
328 - `BetaManagedAgentsAdvisor object`
324 - `version: number`
329325
330 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
326 format: int32
331327
332 - `type: "advisor"`
328 - `BetaManagedAgentsAdvisor object`
333329
334 - `model: string`
330 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
335331
336 The advisor model id.
332 - `type: "advisor"`
337333
334 - `model: string`
335
336 The advisor model id.
337
338 - `BetaManagedAgentsMultiagent20261001 object`
339
340 Resolved multiagent configuration with three members, each enabled or disabled on its own.
341
342 - `type: "multiagent_20261001"`
343
344 - `advisor: BetaManagedAgentsMultiagentAdvisor`
345
346 Whether the session's primary thread can consult an advisor model.
347
348 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
349
350 The session's primary thread can consult `model` mid-turn.
351
352 - `type: "enabled"`
353
354 - `model: string`
355
356 The advisor model id.
357
358 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
359
360 The agent has no advisor.
361
362 - `type: "disabled"`
363
364 - `subagents: BetaManagedAgentsMultiagentSubagents`
365
366 Whether the agent can spawn session threads.
367
368 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
369
370 The agent can spawn session threads.
371
372 - `type: "enabled"`
373
374 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
375
376 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
377
378 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
379
380 The agent can define inline agents.
381
382 - `type: "enabled"`
383
384 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
385
386 The agent cannot define inline agents.
387
388 - `type: "disabled"`
389
390 - `predefined_agents: array of BetaManagedAgentsAgentReference`
391
392 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
393
394 - `type: "agent"`
395
396 - `id: string`
397
398 - `version: number`
399
400 format: int32
401
402 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
403
404 The agent cannot spawn session threads.
405
406 - `type: "disabled"`
407
408 - `workflows: BetaManagedAgentsMultiagentWorkflows`
409
410 Whether the agent can start workflow runs.
411
412 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
413
414 The agent can start workflow runs.
415
416 - `type: "enabled"`
417
418 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
419
420 Whether a run's plan can define inline agents, which are not saved.
421
422 - `predefined_agents: array of BetaManagedAgentsAgentReference`
423
424 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
425
426 - `type: "agent"`
427
428 - `id: string`
429
430 - `version: number`
431
432 format: int32
433
434 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
435
436 The agent cannot start workflow runs.
437
438 - `type: "disabled"`
439
338440 - `name: string`
339441
340442 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 943
841943 "speed": "standard"
842944 },
843945 "multiagent": {
844 "agents": [
845 {
846 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
847 "type": "agent",
848 "version": 1
849 }
850 ],
851 "type": "coordinator"
946 "advisor": {
947 "type": "disabled"
948 },
949 "subagents": {
950 "inline_agents": {
951 "type": "enabled"
952 },
953 "predefined_agents": [
954 {
955 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
956 "type": "agent",
957 "version": 1
958 }
959 ],
960 "type": "enabled"
961 },
962 "type": "multiagent_20261001",
963 "workflows": {
964 "inline_agents": {
965 "type": "enabled"
966 },
967 "predefined_agents": [
968 {
969 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
970 "type": "agent",
971 "version": 1
972 }
973 ],
974 "type": "enabled"
975 }
852976 },
853977 "name": "My First Agent",
854978 "skills": [
api/beta/agents/update Changed · +869 / -646 lines
from line 217
217217
218218 - `"claude-haiku-4-5"`
219219
220 Fastest model with near-frontier intelligence
221
222220 - `"claude-haiku-4-5-20251001"`
223221
224 Fastest model with near-frontier intelligence
225
226222 - `"claude-opus-4-5"`
227223
228224 Powerful intelligence for long-running agents and coding
from line 321
325321
326322 Multiagent orchestration configuration. Full replacement. Omit to preserve; send null to clear.
327323
328 - `type: "coordinator"`
324 - `BetaManagedAgentsMultiagentCoordinatorParams object`
329325
330 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
326 A coordinator topology: the session's primary thread orchestrates work by spawning session threads, each running an agent drawn from the `agents` roster.
331327
332 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
328 - `type: "coordinator"`
333329
334 - `string`
330 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
335331
336 - `BetaManagedAgentsAgentParams object`
332 Agents the coordinator may spawn as session threads. 1–20 entries. Each entry is an agent ID string, a versioned `{"type":"agent","id","version"}` reference, or `{"type":"self"}` to allow recursive self-invocation. Entries must reference distinct agents (after resolving `self` and string forms); at most one `self`. Referenced agents must exist, must not be archived, and must not themselves have `multiagent` set (depth limit 1).
337333
338 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
334 - `string`
339335
340 - `type: "agent"`
336 - `BetaManagedAgentsAgentParams object`
341337
342 - `id: string`
338 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
343339
344 The `agent` ID.
340 - `type: "agent"`
345341
346 minLength: 1, maxLength: 128
342 - `id: string`
347343
348 - `version: optional number`
344 The `agent` ID.
349345
350 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
346 minLength: 1, maxLength: 128
351347
352 format: int32
348 - `version: optional number`
353349
354 - `BetaManagedAgentsMultiagentSelfParams object`
350 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
355351
356 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
352 format: int32
357353
358 - `type: "self"`
354 - `BetaManagedAgentsMultiagentSelfParams object`
359355
360 - `BetaManagedAgentsAdvisorParams object`
356 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
361357
362 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
358 - `type: "self"`
363359
364 - `type: "advisor"`
360 - `BetaManagedAgentsAdvisorParams object`
365361
366 - `model: string`
362 Platform advisor roster entry: a model the session's primary thread may consult mid-turn. At most one per roster; the entry occupies the roster name `anthropic.advisor`.
367363
368 A Claude model id. The model must be permitted as an advisor for this agent's model.
364 - `type: "advisor"`
369365
370 minLength: 1, maxLength: 256
366 - `model: string`
371367
368 A Claude model id. The model must be permitted as an advisor for this agent's model.
369
370 minLength: 1, maxLength: 256
371
372 - `BetaManagedAgentsMultiagent20261001Params object`
373
374 Multiagent configuration with three members, each enabled or disabled on its own. On an update, if the agent's stored `multiagent` also has type `multiagent_20261001`, this configuration is merged into the stored one, level by level, instead of replacing it. A key that the update omits keeps its stored value. A key sent as null takes its default, on create as well, so `"workflows": null` enables workflows. An object sent with a `type` other than the stored one replaces the stored object, and the keys that it omits take their defaults. A `predefined_agents` list that is sent replaces the stored list. Every object that is sent needs its `type`, and an enabled `advisor` needs its `model`. Other validation applies to the merged result.
375
376 - `type: "multiagent_20261001"`
377
378 - `advisor: optional BetaManagedAgentsMultiagentAdvisorParams or null`
379
380 Whether the session's primary thread can consult an advisor model. Defaults to disabled.
381
382 - `BetaManagedAgentsMultiagentAdvisorEnabledParams object`
383
384 The session's primary thread can consult `model` mid-turn.
385
386 - `type: "enabled"`
387
388 - `model: string`
389
390 A Claude model id. The model must be permitted as an advisor for this agent's model.
391
392 minLength: 1, maxLength: 256
393
394 - `BetaManagedAgentsMultiagentAdvisorDisabledParams object`
395
396 The agent has no advisor.
397
398 - `type: "disabled"`
399
400 - `subagents: optional BetaManagedAgentsMultiagentSubagentsParams or null`
401
402 Whether the agent can spawn session threads. Defaults to enabled.
403
404 - `BetaManagedAgentsMultiagentSubagentsEnabledParams object`
405
406 The agent can spawn session threads. Each thread runs a predefined agent, which is a saved agent in `predefined_agents`, or an inline agent, which the agent defines when it spawns the thread and which is not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
407
408 - `type: "enabled"`
409
410 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
411
412 Whether the agent can define inline agents when it spawns session threads. Defaults to enabled.
413
414 - `BetaManagedAgentsMultiagentInlineAgentsEnabledParams object`
415
416 The agent can define inline agents.
417
418 - `type: "enabled"`
419
420 - `BetaManagedAgentsMultiagentInlineAgentsDisabledParams object`
421
422 The agent cannot define inline agents.
423
424 - `type: "disabled"`
425
426 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
427
428 Predefined agents that this agent can spawn as session threads. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `workflows.predefined_agents`, and an agent in one list is not added to the other.
429
430 - `string`
431
432 - `BetaManagedAgentsAgentParams object`
433
434 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
435
436 - `BetaManagedAgentsMultiagentSelfParams object`
437
438 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
439
440 - `BetaManagedAgentsMultiagentSubagentsDisabledParams object`
441
442 The agent cannot spawn session threads.
443
444 - `type: "disabled"`
445
446 - `workflows: optional BetaManagedAgentsMultiagentWorkflowsParams or null`
447
448 Whether the agent can start workflow runs. Defaults to enabled.
449
450 - `BetaManagedAgentsMultiagentWorkflowsEnabledParams object`
451
452 The agent can start workflow runs. Each run follows a plan, a program that the agent writes. A plan can use predefined agents, which are the saved agents in `predefined_agents`, and inline agents, which it defines itself and which are not saved. If `inline_agents` is disabled, `predefined_agents` must name at least one agent.
453
454 - `type: "enabled"`
455
456 - `inline_agents: optional BetaManagedAgentsMultiagentInlineAgentsParams or null`
457
458 Whether a run's plan can define inline agents. Defaults to enabled.
459
460 - `predefined_agents: optional array of BetaManagedAgentsMultiagentPredefinedAgentParams or null`
461
462 Predefined agents that a run's plan can use. At most 20. Defaults to null. Null and an empty list both mean no predefined agents. This list is separate from `subagents.predefined_agents`, and an agent in one list is not added to the other.
463
464 - `string`
465
466 - `BetaManagedAgentsAgentParams object`
467
468 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
469
470 - `BetaManagedAgentsMultiagentSelfParams object`
471
472 Sentinel roster entry meaning "the agent that owns this configuration". Resolved server-side to a concrete agent reference.
473
474 - `BetaManagedAgentsMultiagentWorkflowsDisabledParams object`
475
476 The agent cannot start workflow runs.
477
478 - `type: "disabled"`
479
372480- `name: optional string`
373481
374482 Human-readable name. Must be non-empty. Omit to preserve. Cannot be cleared.
from line 1151
10431151
10441152 - `"claude-haiku-4-5"`
10451153
1046 Fastest model with near-frontier intelligence
1047
10481154 - `"claude-haiku-4-5-20251001"`
10491155
1050 Fastest model with near-frontier intelligence
1051
10521156 - `"claude-opus-4-5"`
10531157
10541158 Powerful intelligence for long-running agents and coding
from line 1225
11211225
11221226 Multiagent orchestration configuration. Null when the agent is single-threaded.
11231227
1124 - `type: "coordinator"`
1228 - `BetaManagedAgentsMultiagentCoordinator object`
11251229
1126 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
1230 Resolved coordinator topology with a concrete agent roster.
11271231
1128 Agents the coordinator may spawn as session threads, each resolved to a specific version.
1232 - `type: "coordinator"`
11291233
1130 - `BetaManagedAgentsAgentReference object`
1234 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
11311235
1132 A resolved agent reference with a concrete version.
1236 Agents the coordinator may spawn as session threads, each resolved to a specific version.
11331237
1134 - `type: "agent"`
1238 - `BetaManagedAgentsAgentReference object`
11351239
1136 - `id: string`
1240 A resolved agent reference with a concrete version.
11371241
1138 - `version: number`
1242 - `type: "agent"`
11391243
1140 format: int32
1244 - `id: string`
11411245
1142 - `BetaManagedAgentsAdvisor object`
1246 - `version: number`
11431247
1144 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
1248 format: int32
11451249
1146 - `type: "advisor"`
1250 - `BetaManagedAgentsAdvisor object`
11471251
1148 - `model: string`
1252 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
11491253
1150 The advisor model id.
1254 - `type: "advisor"`
11511255
1256 - `model: string`
1257
1258 The advisor model id.
1259
1260 - `BetaManagedAgentsMultiagent20261001 object`
1261
1262 Resolved multiagent configuration with three members, each enabled or disabled on its own.
1263
1264 - `type: "multiagent_20261001"`
1265
1266 - `advisor: BetaManagedAgentsMultiagentAdvisor`
1267
1268 Whether the session's primary thread can consult an advisor model.
1269
1270 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
1271
1272 The session's primary thread can consult `model` mid-turn.
1273
1274 - `type: "enabled"`
1275
1276 - `model: string`
1277
1278 The advisor model id.
1279
1280 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
1281
1282 The agent has no advisor.
1283
1284 - `type: "disabled"`
1285
1286 - `subagents: BetaManagedAgentsMultiagentSubagents`
1287
1288 Whether the agent can spawn session threads.
1289
1290 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
1291
1292 The agent can spawn session threads.
1293
1294 - `type: "enabled"`
1295
1296 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
1297
1298 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
1299
1300 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
1301
1302 The agent can define inline agents.
1303
1304 - `type: "enabled"`
1305
1306 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
1307
1308 The agent cannot define inline agents.
1309
1310 - `type: "disabled"`
1311
1312 - `predefined_agents: array of BetaManagedAgentsAgentReference`
1313
1314 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
1315
1316 - `type: "agent"`
1317
1318 - `id: string`
1319
1320 - `version: number`
1321
1322 format: int32
1323
1324 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
1325
1326 The agent cannot spawn session threads.
1327
1328 - `type: "disabled"`
1329
1330 - `workflows: BetaManagedAgentsMultiagentWorkflows`
1331
1332 Whether the agent can start workflow runs.
1333
1334 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
1335
1336 The agent can start workflow runs.
1337
1338 - `type: "enabled"`
1339
1340 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
1341
1342 Whether a run's plan can define inline agents, which are not saved.
1343
1344 - `predefined_agents: array of BetaManagedAgentsAgentReference`
1345
1346 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
1347
1348 - `type: "agent"`
1349
1350 - `id: string`
1351
1352 - `version: number`
1353
1354 format: int32
1355
1356 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
1357
1358 The agent cannot start workflow runs.
1359
1360 - `type: "disabled"`
1361
11521362 - `name: string`
11531363
11541364 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 1839
16291839 -H "X-Api-Key: $ANTHROPIC_API_KEY" \
16301840 -d '{
16311841 "description": "updated",
1632 "multiagent": {
1633 "agents": [
1634 "agent_011CZkYqphY8vELVzwCUpqiQ",
1635 {
1636 "type": "self"
1637 }
1638 ],
1639 "type": "coordinator"
1640 },
16411842 "system": "You are a general-purpose agent that can research, write code, run commands, and use connected tools to complete the user'\''s task end to end.",
16421843 "version": 1
16431844 }'
from line 1871
16701871 "speed": "standard"
16711872 },
16721873 "multiagent": {
1673 "agents": [
1674 {
1675 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1676 "type": "agent",
1677 "version": 1
1678 }
1679 ],
1680 "type": "coordinator"
1874 "advisor": {
1875 "type": "disabled"
1876 },
1877 "subagents": {
1878 "inline_agents": {
1879 "type": "enabled"
1880 },
1881 "predefined_agents": [
1882 {
1883 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1884 "type": "agent",
1885 "version": 1
1886 }
1887 ],
1888 "type": "enabled"
1889 },
1890 "type": "multiagent_20261001",
1891 "workflows": {
1892 "inline_agents": {
1893 "type": "enabled"
1894 },
1895 "predefined_agents": [
1896 {
1897 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
1898 "type": "agent",
1899 "version": 1
1900 }
1901 ],
1902 "type": "enabled"
1903 }
16811904 },
16821905 "name": "My First Agent",
16831906 "skills": [
api/beta/agents/versions Changed · +163 / -39 lines
from line 235
235235
236236 - `"claude-haiku-4-5"`
237237
238 Fastest model with near-frontier intelligence
239
240238 - `"claude-haiku-4-5-20251001"`
241239
242 Fastest model with near-frontier intelligence
243
244240 - `"claude-opus-4-5"`
245241
246242 Powerful intelligence for long-running agents and coding
from line 309
313309
314310 Multiagent orchestration configuration. Null when the agent is single-threaded.
315311
316 - `type: "coordinator"`
312 - `BetaManagedAgentsMultiagentCoordinator object`
317313
318 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
314 Resolved coordinator topology with a concrete agent roster.
319315
320 Agents the coordinator may spawn as session threads, each resolved to a specific version.
316 - `type: "coordinator"`
321317
322 - `BetaManagedAgentsAgentReference object`
318 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
323319
324 A resolved agent reference with a concrete version.
320 Agents the coordinator may spawn as session threads, each resolved to a specific version.
325321
326 - `type: "agent"`
322 - `BetaManagedAgentsAgentReference object`
327323
328 - `id: string`
324 A resolved agent reference with a concrete version.
329325
330 - `version: number`
326 - `type: "agent"`
331327
332 format: int32
328 - `id: string`
333329
334 - `BetaManagedAgentsAdvisor object`
330 - `version: number`
335331
336 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
332 format: int32
337333
338 - `type: "advisor"`
334 - `BetaManagedAgentsAdvisor object`
339335
340 - `model: string`
336 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
341337
342 The advisor model id.
338 - `type: "advisor"`
343339
340 - `model: string`
341
342 The advisor model id.
343
344 - `BetaManagedAgentsMultiagent20261001 object`
345
346 Resolved multiagent configuration with three members, each enabled or disabled on its own.
347
348 - `type: "multiagent_20261001"`
349
350 - `advisor: BetaManagedAgentsMultiagentAdvisor`
351
352 Whether the session's primary thread can consult an advisor model.
353
354 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
355
356 The session's primary thread can consult `model` mid-turn.
357
358 - `type: "enabled"`
359
360 - `model: string`
361
362 The advisor model id.
363
364 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
365
366 The agent has no advisor.
367
368 - `type: "disabled"`
369
370 - `subagents: BetaManagedAgentsMultiagentSubagents`
371
372 Whether the agent can spawn session threads.
373
374 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
375
376 The agent can spawn session threads.
377
378 - `type: "enabled"`
379
380 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
381
382 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
383
384 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
385
386 The agent can define inline agents.
387
388 - `type: "enabled"`
389
390 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
391
392 The agent cannot define inline agents.
393
394 - `type: "disabled"`
395
396 - `predefined_agents: array of BetaManagedAgentsAgentReference`
397
398 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
399
400 - `type: "agent"`
401
402 - `id: string`
403
404 - `version: number`
405
406 format: int32
407
408 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
409
410 The agent cannot spawn session threads.
411
412 - `type: "disabled"`
413
414 - `workflows: BetaManagedAgentsMultiagentWorkflows`
415
416 Whether the agent can start workflow runs.
417
418 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
419
420 The agent can start workflow runs.
421
422 - `type: "enabled"`
423
424 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
425
426 Whether a run's plan can define inline agents, which are not saved.
427
428 - `predefined_agents: array of BetaManagedAgentsAgentReference`
429
430 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
431
432 - `type: "agent"`
433
434 - `id: string`
435
436 - `version: number`
437
438 format: int32
439
440 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
441
442 The agent cannot start workflow runs.
443
444 - `type: "disabled"`
445
344446 - `name: string`
345447
346448 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 955
853955 "speed": "standard"
854956 },
855957 "multiagent": {
856 "agents": [
857 {
858 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
859 "type": "agent",
860 "version": 1
861 }
862 ],
863 "type": "coordinator"
958 "advisor": {
959 "type": "disabled"
960 },
961 "subagents": {
962 "inline_agents": {
963 "type": "enabled"
964 },
965 "predefined_agents": [
966 {
967 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
968 "type": "agent",
969 "version": 1
970 }
971 ],
972 "type": "enabled"
973 },
974 "type": "multiagent_20261001",
975 "workflows": {
976 "inline_agents": {
977 "type": "enabled"
978 },
979 "predefined_agents": [
980 {
981 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
982 "type": "agent",
983 "version": 1
984 }
985 ],
986 "type": "enabled"
987 }
864988 },
865989 "name": "My First Agent",
866990 "skills": [
api/beta/agents/versions/list Changed · +163 / -39 lines
from line 233
233233
234234 - `"claude-haiku-4-5"`
235235
236 Fastest model with near-frontier intelligence
237
238236 - `"claude-haiku-4-5-20251001"`
239237
240 Fastest model with near-frontier intelligence
241
242238 - `"claude-opus-4-5"`
243239
244240 Powerful intelligence for long-running agents and coding
from line 307
311307
312308 Multiagent orchestration configuration. Null when the agent is single-threaded.
313309
314 - `type: "coordinator"`
310 - `BetaManagedAgentsMultiagentCoordinator object`
315311
316 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
312 Resolved coordinator topology with a concrete agent roster.
317313
318 Agents the coordinator may spawn as session threads, each resolved to a specific version.
314 - `type: "coordinator"`
319315
320 - `BetaManagedAgentsAgentReference object`
316 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
321317
322 A resolved agent reference with a concrete version.
318 Agents the coordinator may spawn as session threads, each resolved to a specific version.
323319
324 - `type: "agent"`
320 - `BetaManagedAgentsAgentReference object`
325321
326 - `id: string`
322 A resolved agent reference with a concrete version.
327323
328 - `version: number`
324 - `type: "agent"`
329325
330 format: int32
326 - `id: string`
331327
332 - `BetaManagedAgentsAdvisor object`
328 - `version: number`
333329
334 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
330 format: int32
335331
336 - `type: "advisor"`
332 - `BetaManagedAgentsAdvisor object`
337333
338 - `model: string`
334 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
339335
340 The advisor model id.
336 - `type: "advisor"`
341337
338 - `model: string`
339
340 The advisor model id.
341
342 - `BetaManagedAgentsMultiagent20261001 object`
343
344 Resolved multiagent configuration with three members, each enabled or disabled on its own.
345
346 - `type: "multiagent_20261001"`
347
348 - `advisor: BetaManagedAgentsMultiagentAdvisor`
349
350 Whether the session's primary thread can consult an advisor model.
351
352 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
353
354 The session's primary thread can consult `model` mid-turn.
355
356 - `type: "enabled"`
357
358 - `model: string`
359
360 The advisor model id.
361
362 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
363
364 The agent has no advisor.
365
366 - `type: "disabled"`
367
368 - `subagents: BetaManagedAgentsMultiagentSubagents`
369
370 Whether the agent can spawn session threads.
371
372 - `BetaManagedAgentsMultiagentSubagentsEnabled object`
373
374 The agent can spawn session threads.
375
376 - `type: "enabled"`
377
378 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
379
380 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
381
382 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
383
384 The agent can define inline agents.
385
386 - `type: "enabled"`
387
388 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
389
390 The agent cannot define inline agents.
391
392 - `type: "disabled"`
393
394 - `predefined_agents: array of BetaManagedAgentsAgentReference`
395
396 Predefined agents, which are saved agents that this agent can spawn as session threads, each resolved to a specific version.
397
398 - `type: "agent"`
399
400 - `id: string`
401
402 - `version: number`
403
404 format: int32
405
406 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
407
408 The agent cannot spawn session threads.
409
410 - `type: "disabled"`
411
412 - `workflows: BetaManagedAgentsMultiagentWorkflows`
413
414 Whether the agent can start workflow runs.
415
416 - `BetaManagedAgentsMultiagentWorkflowsEnabled object`
417
418 The agent can start workflow runs.
419
420 - `type: "enabled"`
421
422 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
423
424 Whether a run's plan can define inline agents, which are not saved.
425
426 - `predefined_agents: array of BetaManagedAgentsAgentReference`
427
428 Predefined agents, which are saved agents that a run's plan can use, each resolved to a specific version.
429
430 - `type: "agent"`
431
432 - `id: string`
433
434 - `version: number`
435
436 format: int32
437
438 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
439
440 The agent cannot start workflow runs.
441
442 - `type: "disabled"`
443
342444 - `name: string`
343445
344446 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 953
851953 "speed": "standard"
852954 },
853955 "multiagent": {
854 "agents": [
855 {
856 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
857 "type": "agent",
858 "version": 1
859 }
860 ],
861 "type": "coordinator"
956 "advisor": {
957 "type": "disabled"
958 },
959 "subagents": {
960 "inline_agents": {
961 "type": "enabled"
962 },
963 "predefined_agents": [
964 {
965 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
966 "type": "agent",
967 "version": 1
968 }
969 ],
970 "type": "enabled"
971 },
972 "type": "multiagent_20261001",
973 "workflows": {
974 "inline_agents": {
975 "type": "enabled"
976 },
977 "predefined_agents": [
978 {
979 "id": "agent_011CZkYqphY8vELVzwCUpqiQ",
980 "type": "agent",
981 "version": 1
982 }
983 ],
984 "type": "enabled"
985 }
862986 },
863987 "name": "My First Agent",
864988 "skills": [
api/beta/messages Changed · +35 / -182 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
from line 161
161161
162162 Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages.
163163
164 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response.
164 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. This is called prefill. On models that don't support prefill, creating a message that ends with a partial `assistant` response returns a 400 error. See [Prefill not supported](https://platform.claude.com/docs/en/api/errors#prefill-not-supported).
165165
166166 Example with a single `user` message:
167167
from line 179
179179 ]
180180 ```
181181
182 Example with a partially-filled response from Claude:
182 Example with a partially-filled response from Claude, for models that support prefill:
183183
184184 ```json
185185 [
from line 3166
31663166
31673167 - `"claude-haiku-4-5"`
31683168
3169 Fastest model with near-frontier intelligence
3170
31713169 - `"claude-haiku-4-5-20251001"`
31723170
3173 Fastest model with near-frontier intelligence
3174
31753171 - `"claude-opus-4-5"`
31763172
31773173 Powerful intelligence for long-running agents and coding
from line 3762
37663762
37673763- `fallbacks: optional BetaFallbacksParam or null`
37683764
3769 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. The string "default" requests the requested model's server-defined default fallback configuration.
3765 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. Some models don't support fallbacks; on those models, a list of fallback models returns a 400 error. See [Server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#server-side-fallback). The string "default" requests the requested model's server-defined default fallback configuration. On a model that doesn't support fallbacks, the request runs on the requested model alone, so a declined request stays declined.
37703766
37713767 - `array of BetaFallbackParam`
37723768
from line 3860
38643860
38653861 - `display: optional "summarized" or "omitted" or "updates" or null`
38663862
3867 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3863 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
38683864
38693865 - `"summarized"`
38703866
from line 3886
38903886
38913887 - `display: optional "summarized" or "omitted" or "updates" or null`
38923888
3893 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3889 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
38943890
38953891 - `"summarized"`
38963892
from line 3992
39963992
39973993- `thinking: optional BetaThinkingConfigParam`
39983994
3999 Configuration for enabling Claude's extended thinking.
3995 Configuration for Claude's thinking.
40003996
4001 When enabled, responses include `thinking` content blocks showing Claude's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit.
3997 With `{"type": "adaptive"}`, Claude decides when and how much to think. With `{"type": "enabled"}` (manual extended thinking), you set a `budget_tokens` of at least 1,024. Thinking tokens count toward your `max_tokens` limit.
40023998
4003 See [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for details.
3999 Which `type` values are accepted, and what happens when you omit `thinking`, depend on the model. See [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking#configuring-thinking) for each model's behavior.
40044000
40054001 - `BetaThinkingConfigEnabled object`
40064002
from line 7032
70367032
70377033 - `"claude-haiku-4-5"`
70387034
7039 Fastest model with near-frontier intelligence
7040
70417035 - `"claude-haiku-4-5-20251001"`
70427036
7043 Fastest model with near-frontier intelligence
7044
70457037 - `"claude-opus-4-5"`
70467038
70477039 Powerful intelligence for long-running agents and coding
from line 7174
71827174
71837175 If true, tool will not be included in initial system prompt. Only loaded when returned via tool_reference from tool search.
71847176
7185 - `strict: optional boolean`
7186
7187 When true, guarantees schema validation on tool names and inputs
7188
7189 - `BetaMCPToolset object`
7190
7191 Configuration for a group of tools from an MCP server.
7192
7193 Allows configuring enabled status and defer_loading for all tools
7194 from an MCP server, with optional per-tool overrides.
7195
7196 - `type: "mcp_toolset"`
7197
7198 - `mcp_server_name: string`
7199
7200 Name of the MCP server to configure tools for
7201
7202 minLength: 1, maxLength: 255
7203
7204 - `cache_control: optional BetaCacheControlEphemeral or null`
7205
7206 Create a cache control breakpoint at this content block.
7207
7208
7177 - `strict
api/beta/messages/batches Changed · +8 / -32 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
from line 181
181181
182182 Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages.
183183
184 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response.
184 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. This is called prefill. On models that don't support prefill, creating a message that ends with a partial `assistant` response returns a 400 error. See [Prefill not supported](https://platform.claude.com/docs/en/api/errors#prefill-not-supported).
185185
186186 Example with a single `user` message:
187187
from line 199
199199 ]
200200 ```
201201
202 Example with a partially-filled response from Claude:
202 Example with a partially-filled response from Claude, for models that support prefill:
203203
204204 ```json
205205 [
from line 3186
31863186
31873187 - `"claude-haiku-4-5"`
31883188
3189 Fastest model with near-frontier intelligence
3190
31913189 - `"claude-haiku-4-5-20251001"`
31923190
3193 Fastest model with near-frontier intelligence
3194
31953191 - `"claude-opus-4-5"`
31963192
31973193 Powerful intelligence for long-running agents and coding
from line 3782
37863782
37873783 - `fallbacks: optional BetaFallbacksParam or null`
37883784
3789 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. The string "default" requests the requested model's server-defined default fallback configuration.
3785 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. Some models don't support fallbacks; on those models, a list of fallback models returns a 400 error. See [Server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#server-side-fallback). The string "default" requests the requested model's server-defined default fallback configuration. On a model that doesn't support fallbacks, the request runs on the requested model alone, so a declined request stays declined.
37903786
37913787 - `array of BetaFallbackParam`
37923788
from line 3880
38843880
38853881 - `display: optional "summarized" or "omitted" or "updates" or null`
38863882
3887 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3883 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
38883884
38893885 - `"summarized"`
38903886
from line 3906
39103906
39113907 - `display: optional "summarized" or "omitted" or "updates" or null`
39123908
3913 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3909 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
39143910
39153911 - `"summarized"`
39163912
from line 4012
40164012
40174013 - `thinking: optional BetaThinkingConfigParam`
40184014
4019 Configuration for enabling Claude's extended thinking.
4015 Configuration for Claude's thinking.
40204016
4021 When enabled, responses include `thinking` content blocks showing Claude's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit.
4017 With `{"type": "adaptive"}`, Claude decides when and how much to think. With `{"type": "enabled"}` (manual extended thinking), you set a `budget_tokens` of at least 1,024. Thinking tokens count toward your `max_tokens` limit.
40224018
4023 See [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for details.
4019 Which `type` values are accepted, and what happens when you omit `thinking`, depend on the model. See [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking#configuring-thinking) for each model's behavior.
40244020
40254021 - `BetaThinkingConfigEnabled object`
40264022
from line 7338
73427338
73437339 When true, guarantees schema validation on tool names and inputs
73447340
7345 - `BetaToolComputerUse20250124 object`
7346
7347 - `type: "computer_20250124"`
7348
7349 - `display_height_px: number`
7350
7351 The height of the display in pixels.
7352
7353 minimum: 1
7354
7355 - `display_width_px: number`
7356
7357 The width of the display in pixels.
7358
7359 minimum: 1
7360
7361 - `name: "computer"`
7362
7363 Name of the tool.
7364
7365 This is how the tool will be called by the model and in `tool_use` blocks.
7366
7367 - `allowed_callers: optional array of "direct" or "code_execution_20250825" or "code_execution_20260120" or "code_execution_20260521"`
7368
7369 - `"direct"`
7370
7371 - `"code_execution_20250825"`
7372
7373 - `"code_execution_2
7341 - `BetaToolComputerUse20
api/beta/messages/batches/create Changed · +10 / -14 lines
from line 179
179179
180180 Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages.
181181
182 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response.
182 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. This is called prefill. On models that don't support prefill, creating a message that ends with a partial `assistant` response returns a 400 error. See [Prefill not supported](https://platform.claude.com/docs/en/api/errors#prefill-not-supported).
183183
184184 Example with a single `user` message:
185185
from line 197
197197 ]
198198 ```
199199
200 Example with a partially-filled response from Claude:
200 Example with a partially-filled response from Claude, for models that support prefill:
201201
202202 ```json
203203 [
from line 3184
31843184
31853185 - `"claude-haiku-4-5"`
31863186
3187 Fastest model with near-frontier intelligence
3188
31893187 - `"claude-haiku-4-5-20251001"`
31903188
3191 Fastest model with near-frontier intelligence
3192
31933189 - `"claude-opus-4-5"`
31943190
31953191 Powerful intelligence for long-running agents and coding
from line 3780
37843780
37853781 - `fallbacks: optional BetaFallbacksParam or null`
37863782
3787 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. The string "default" requests the requested model's server-defined default fallback configuration.
3783 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. Some models don't support fallbacks; on those models, a list of fallback models returns a 400 error. See [Server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#server-side-fallback). The string "default" requests the requested model's server-defined default fallback configuration. On a model that doesn't support fallbacks, the request runs on the requested model alone, so a declined request stays declined.
37883784
37893785 - `array of BetaFallbackParam`
37903786
from line 3878
38823878
38833879 - `display: optional "summarized" or "omitted" or "updates" or null`
38843880
3885 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3881 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
38863882
38873883 - `"summarized"`
38883884
from line 3904
39083904
39093905 - `display: optional "summarized" or "omitted" or "updates" or null`
39103906
3911 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3907 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
39123908
39133909 - `"summarized"`
39143910
from line 4010
40144010
40154011 - `thinking: optional BetaThinkingConfigParam`
40164012
4017 Configuration for enabling Claude's extended thinking.
4013 Configuration for Claude's thinking.
40184014
4019 When enabled, responses include `thinking` content blocks showing Claude's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit.
4015 With `{"type": "adaptive"}`, Claude decides when and how much to think. With `{"type": "enabled"}` (manual extended thinking), you set a `budget_tokens` of at least 1,024. Thinking tokens count toward your `max_tokens` limit.
40204016
4021 See [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for details.
4017 Which `type` values are accepted, and what happens when you omit `thinking`, depend on the model. See [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking#configuring-thinking) for each model's behavior.
40224018
40234019 - `BetaThinkingConfigEnabled object`
40244020
api/beta/messages/batches/results Changed · +0 / -5 lines
from line 2948
29482948
29492949 - `"claude-haiku-4-5"`
29502950
2951 Fastest model with near-frontier intelligence
2952
29532951 - `"claude-haiku-4-5-20251001"`
29542952
2955 Fastest model with near-frontier intelligence
2956
29572953 - `"claude-opus-4-5"`
29582954
29592955 Powerful intelligence for long-running agents and coding
from line 3650
36543650
36553651 Each entry represents one sampling iteration, with its own input/output token counts and cache statistics, discriminated by `type`. For `message` entries (model sampling iterations, such as the turns of a server-side tool use loop), this allows you to:
36563652
3657 - Determine which iterations exceeded long context thresholds (>=200k tokens)
36583653 - Calculate the context window size from the last `message` entry
36593654 - Understand token accumulation across server-side tool use loops
36603655
api/beta/messages/count_tokens Changed · +9 / -13 lines
from line 147
147147
148148 Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages.
149149
150 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response.
150 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. This is called prefill. On models that don't support prefill, creating a message that ends with a partial `assistant` response returns a 400 error. See [Prefill not supported](https://platform.claude.com/docs/en/api/errors#prefill-not-supported).
151151
152152 Example with a single `user` message:
153153
from line 165
165165 ]
166166 ```
167167
168 Example with a partially-filled response from Claude:
168 Example with a partially-filled response from Claude, for models that support prefill:
169169
170170 ```json
171171 [
from line 3152
31523152
31533153 - `"claude-haiku-4-5"`
31543154
3155 Fastest model with near-frontier intelligence
3156
31573155 - `"claude-haiku-4-5-20251001"`
31583156
3159 Fastest model with near-frontier intelligence
3160
31613157 - `"claude-opus-4-5"`
31623158
31633159 Powerful intelligence for long-running agents and coding
from line 3749
37533749
37543750- `thinking: optional BetaThinkingConfigParam`
37553751
3756 Configuration for enabling Claude's extended thinking.
3752 Configuration for Claude's thinking.
37573753
3758 When enabled, responses include `thinking` content blocks showing Claude's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit.
3754 With `{"type": "adaptive"}`, Claude decides when and how much to think. With `{"type": "enabled"}` (manual extended thinking), you set a `budget_tokens` of at least 1,024. Thinking tokens count toward your `max_tokens` limit.
37593755
3760 See [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for details.
3756 Which `type` values are accepted, and what happens when you omit `thinking`, depend on the model. See [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking#configuring-thinking) for each model's behavior.
37613757
37623758 - `BetaThinkingConfigEnabled object`
37633759
from line 3783
37873783
37883784 - `display: optional "summarized" or "omitted" or "updates" or null`
37893785
3790 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3786 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
37913787
37923788 - `"summarized"`
37933789
from line 3809
38133809
38143810 - `display: optional "summarized" or "omitted" or "updates" or null`
38153811
3816 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3812 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
38173813
38183814 - `"summarized"`
38193815
api/beta/messages/create Changed · +10 / -20 lines
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
from line 159
159159
160160 Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages.
161161
162 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response.
162 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. This is called prefill. On models that don't support prefill, creating a message that ends with a partial `assistant` response returns a 400 error. See [Prefill not supported](https://platform.claude.com/docs/en/api/errors#prefill-not-supported).
163163
164164 Example with a single `user` message:
165165
from line 177
177177 ]
178178 ```
179179
180 Example with a partially-filled response from Claude:
180 Example with a partially-filled response from Claude, for models that support prefill:
181181
182182 ```json
183183 [
from line 3164
31643164
31653165 - `"claude-haiku-4-5"`
31663166
3167 Fastest model with near-frontier intelligence
3168
31693167 - `"claude-haiku-4-5-20251001"`
31703168
3171 Fastest model with near-frontier intelligence
3172
31733169 - `"claude-opus-4-5"`
31743170
31753171 Powerful intelligence for long-running agents and coding
from line 3760
37643760
37653761- `fallbacks: optional BetaFallbacksParam or null`
37663762
3767 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. The string "default" requests the requested model's server-defined default fallback configuration.
3763 Opt-in server-side retry on one or more substitute models when the requested model declines for policy reasons. Tried in order: if the first entry also declines, the second is tried, and so on. Some models don't support fallbacks; on those models, a list of fallback models returns a 400 error. See [Server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#server-side-fallback). The string "default" requests the requested model's server-defined default fallback configuration. On a model that doesn't support fallbacks, the request runs on the requested model alone, so a declined request stays declined.
37683764
37693765 - `array of BetaFallbackParam`
37703766
from line 3858
38623858
38633859 - `display: optional "summarized" or "omitted" or "updates" or null`
38643860
3865 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3861 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
38663862
38673863 - `"summarized"`
38683864
from line 3884
38883884
38893885 - `display: optional "summarized" or "omitted" or "updates" or null`
38903886
3891 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
3887 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
38923888
38933889 - `"summarized"`
38943890
from line 3990
39943990
39953991- `thinking: optional BetaThinkingConfigParam`
39963992
3997 Configuration for enabling Claude's extended thinking.
3993 Configuration for Claude's thinking.
39983994
3999 When enabled, responses include `thinking` content blocks showing Claude's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit.
3995 With `{"type": "adaptive"}`, Claude decides when and how much to think. With `{"type": "enabled"}` (manual extended thinking), you set a `budget_tokens` of at least 1,024. Thinking tokens count toward your `max_tokens` limit.
40003996
4001 See [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for details.
3997 Which `type` values are accepted, and what happens when you omit `thinking`, depend on the model. See [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking#configuring-thinking) for each model's behavior.
40023998
40033999 - `BetaThinkingConfigEnabled object`
40044000
from line 7030
70347030
70357031 - `"claude-haiku-4-5"`
70367032
7037 Fastest model with near-frontier intelligence
7038
70397033 - `"claude-haiku-4-5-20251001"`
70407034
7041 Fastest model with near-frontier intelligence
7042
70437035 - `"claude-opus-4-5"`
70447036
70457037 Powerful intelligence for long-running agents and coding
from line 7172
71807172
71817173 If true, tool will not be included in initial system prompt. Only loaded when returned via tool_reference from tool search.
71827174
7183 - `strict: optional boolean`
7184
7185 When true, guarantees schema validation on tool names and inputs
7186
7187 - `BetaMCPToolset object`
7188
7189 Configuration for a group of tools from an MCP server.
7190
7191 Allows configuring enabled status and defer_loading for all tools
7192 from an MCP server, with optional per-tool overrides.
7193
7194 - `type: "mcp_toolset"`
7195
7196 - `mcp_server_name: string`
7197
7198 Name of the MCP server to configure tools for
7199
7200 minLength: 1, maxLength: 255
7201
7202 - `cache_control: optional BetaCacheControlEphemeral or null`
7203
7204 Create a cache control breakpoint at this content block.
7205
7206
7175 - `strict:
api/beta/sessions Changed · +10535 / -3492 lines
### Beta Managed Agents Session Multiagent ### Beta Managed Agents Session Multiagent Subagents ### Beta Managed Agents Session Multiagent Subagents Enabled ### Beta Managed Agents Session Multiagent Workflows ### Beta Managed Agents Session Multiagent Workflows Enabled ### Beta Managed Agents Session Multiagent20261001
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
Nothing in the body moved in this read. What changed is above.
api/beta/sessions/archive Changed · +571 / -437 lines
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 213
213213
214214 - `"claude-haiku-4-5"`
215215
216 Fastest model with near-frontier intelligence
217
218216 - `"claude-haiku-4-5-20251001"`
219217
220 Fastest model with near-frontier intelligence
221
222218 - `"claude-opus-4-5"`
223219
224220 Powerful intelligence for long-running agents and coding
from line 283
287283
288284 - `"fast"`
289285
290 - `multiagent: BetaManagedAgentsSessionMultiagentCoordinator or null`
286 - `multiagent: BetaManagedAgentsSessionMultiagent or null`
291287
292288 Resolved multiagent orchestration configuration. Null when the agent is single-threaded.
293289
294 - `type: "coordinator"`
295
296 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
297
298 Full `agent` definitions the coordinator may spawn as session threads.
299
300 - `BetaManagedAgentsSessionThreadAgent object`
301
302 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
303
304 - `type: "agent"`
305
306 - `id: string`
307
308 - `description: string or null`
309
310 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
311
312 - `type: "url"`
290 - `BetaManagedAgentsSessionMultiagentCoordinator object`
291
292 Resolved coordinator topology with full agent definitions for each roster member.
293
294 - `type: "coordinator"`
295
296 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
297
298 Full `agent` definitions the coordinator may spawn as session threads.
299
300 - `BetaManagedAgentsSessionThreadAgent object`
301
302 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
303
304 - `type: "agent"`
305
306 - `id: string`
307
308 - `description: string or null`
309
310 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
311
312 - `type: "url"`
313
314 - `name: string`
315
316 - `url: string`
317
318 - `model: BetaManagedAgentsModelConfig`
319
320 Model identifier and configuration.
313321
314322 - `name: string`
315323
316 - `url: string`
317
318 - `model: BetaManagedAgentsModelConfig`
319
320 Model identifier and configuration.
321
322 - `name: string`
323
324 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
325
326 - `BetaManagedAgentsAnthropicSkill object`
327
328 A resolved Anthropic-managed skill.
329
330 - `type: "anthropic"`
331
332 - `skill_id: string`
333
334 - `version: string`
335
336 - `BetaManagedAgentsCustomSkill object`
337
338 A resolved user-created custom skill.
339
340 - `type: "custom"`
341
342 - `skill_id: string`
343
344 - `version: string`
345
346 - `system: string or null`
347
348 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
349
350 - `BetaManagedAgentsAgentToolset20260401 object`
351
352 - `type: "agent_toolset_20260401"`
353
354 - `configs: array of BetaManagedAgentsAgentToolConfig`
355
356 - `BetaManagedAgentsBashToolConfig object`
357
358 Configuration for the bash tool.
359
360 - `type: "bash"`
324 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
325
326 - `BetaManagedAgentsAnthropicSkill object`
327
328 A resolved Anthropic-managed skill.
329
330 - `type: "anthropic"`
331
332 - `skill_id: string`
333
334 - `version: string`
335
336 - `BetaManagedAgentsCustomSkill object`
337
338 A resolved user-created custom skill.
339
340 - `type: "custom"`
341
342 - `skill_id: string`
343
344 - `version: string`
345
346 - `system: string or null`
347
348 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
349
350 - `BetaManagedAgentsAgentToolset20260401 object`
351
352 - `type: "agent_toolset_20260401"`
353
354 - `configs: array of BetaManagedAgentsAgentToolConfig`
355
356 - `BetaManagedAgentsBashToolConfig object`
357
358 Configuration for the bash tool.
359
360 - `type: "bash"`
361
362 - `enabled: boolean`
363
364 - `name: "bash"`
365
366 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
367
368 Permission policy for tool execution.
369
370 - `BetaManagedAgentsAlwaysAllowPolicy object`
371
372 Tool calls are automatically approved without user confirmation.
373
374 - `type: "always_allow"`
375
376 - `BetaManagedAgentsAlwaysAskPolicy object`
377
378 Tool calls require user confirmation before execution.
379
380 - `type: "always_ask"`
381
382 - `BetaManagedAgentsAutoPolicy object`
383
384 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
385
386 - `type: "auto"`
387
388 - `BetaManagedAgentsEditToolConfig object`
389
390 Configuration for the edit tool.
391
392 - `type: "edit"`
393
394 - `enabled: boolean`
395
396 - `name: "edit"`
397
398 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
399
400 Permission policy for tool execution.
401
402 - `BetaManagedAgentsAlwaysAllowPolicy object`
403
404 Tool calls are automatically approved without user confirmation.
405
406 - `BetaManagedAgentsAlwaysAskPolicy object`
407
408 Tool calls require user confirmation before execution.
409
410 - `BetaManagedAgentsAutoPolicy object`
411
412 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
413
414 - `BetaManagedAgentsReadToolConfig object`
415
416 Configuration for the read tool.
417
418 - `type: "read"`
419
420 - `enabled: boolean`
421
422 - `name: "read"`
423
424 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
425
426 Permission policy for tool execution.
427
428 - `BetaManagedAgentsAlwaysAllowPolicy object`
429
430 Tool calls are automatically approved without user confirmation.
431
432 - `BetaManagedAgentsAlwaysAskPolicy object`
433
434 Tool calls require user confirmation before execution.
435
436 - `BetaManagedAgentsAutoPolicy object`
437
438 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
439
440 - `BetaManagedAgentsWriteToolConfig object`
441
442 Configuration for the write tool.
443
444 - `type: "write"`
445
446 - `enabled: boolean`
447
448 - `name: "write"`
449
450 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
451
452 Permission policy for tool execution.
453
454 - `BetaManagedAgentsAlwaysAllowPolicy object`
455
456 Tool calls are automatically approved without user confirmation.
457
458 - `BetaManagedAgentsAlwaysAskPolicy object`
459
460 Tool calls require user confirmation before execution.
461
462 - `BetaManagedAgentsAutoPolicy object`
463
464 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
465
466 - `BetaManagedAgentsGlobToolConfig object`
467
468 Configuration for the glob tool.
469
470 - `type: "glob"`
471
472 - `enabled: boolean`
473
474 - `name: "glob"`
475
476 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
477
478 Permission policy for tool execution.
479
480 - `BetaManagedAgentsAlwaysAllowPolicy object`
481
482 Tool calls are automatically approved without user confirmation.
483
484 - `BetaManagedAgentsAlwaysAskPolicy object`
485
486 Tool calls require user confirmation before execution.
487
488 - `BetaManagedAgentsAutoPolicy object`
489
490 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
491
492 - `BetaManagedAgentsGrepToolConfig object`
493
494 Configuration for the grep tool.
495
496 - `type: "grep"`
497
498 - `enabled: boolean`
499
500 - `name: "grep"`
501
502 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
503
504 Permission policy for tool execution.
505
506 - `BetaManagedAgentsAlwaysAllowPolicy object`
507
508 Tool calls are automatically approved without user confirmation.
509
510 - `BetaManagedAgentsAlwaysAskPolicy object`
511
512 Tool calls require user confirmation before execution.
513
514 - `BetaManagedAgentsAutoPolicy object`
515
516 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
517
518 - `BetaManagedAgentsWebFetchToolConfig object`
519
520 Configuration for the web_fetch tool.
521
522 - `type: "web_fetch"`
523
524 - `enabled: boolean`
525
526 - `name: "web_fetch"`
527
528 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
529
530 Permission policy for tool execution.
531
532 - `BetaManagedAgentsAlwaysAllowPolicy object`
533
534 Tool calls are automatically approved without user confirmation.
535
536 - `BetaManagedAgentsAlwaysAskPolicy object`
537
538 Tool calls require user confirmation before execution.
539
540 - `BetaManagedAgentsAutoPolicy object`
541
542 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
543
544 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
545
546 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
547
548 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
549
550 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
551
552 - `BetaManagedAgentsWebFetchURLSourceAll object`
553
554 Every URL from this source may be fetched. This is the default.
555
556 - `type: "all"`
557
558 - `BetaManagedAgentsWebFetchURLSourceNone object`
559
560 This source contributes no URLs that may be fetched.
561
562 - `type: "none"`
563
564 - `BetaManagedAgentsWebFetchURLSourceOnly object`
565
566 Only the named tools' results contribute URLs that may be fetched.
567
568 - `type: "only"`
569
570 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
571
572 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
573
574 - `type: "tool_reference"`
575
576 Must be "tool_reference".
577
578 - `name: string`
579
580 Name of the tool. Compared exactly, so upper and lower case letters are different.
581
582 minLength: 1, maxLength: 128
583
584 - `BetaManagedAgentsWebFetchURLSourceExcept object`
585
586 Every tool's results contribute URLs that may be fetched, except the named tools' results.
587
588 - `type: "except"`
589
590 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
591
592 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
593
594 - `type: "tool_reference"`
595
596 Must be "tool_reference".
597
598 - `name: string`
599
600 Name of the tool. Compared exactly, so upper and lower case letters are different.
601
602 minLength: 1, maxLength: 128
603
604 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
605
606 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
607
608 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
609
610 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
611
612 - `BetaManagedAgentsWebFetchURLSourceAll object`
613
614 Every URL from this source may be fetched. This is the default.
615
616 - `BetaManagedAgentsWebFetchURLSourceNone object`
617
618 This source contributes no URLs that may be fetched.
619
620 - `allowed_domains: optional array of string`
621
622 - `blocked_domains: optional array of string`
623
624 - `max_content_tokens: optional number or null`
625
626 format: int32
627
628 - `BetaManagedAgentsWebSearchToolConfig object`
629
630 Configuration for the web_search tool.
631
632 - `type: "web_search"`
633
634 - `enabled: boolean`
635
636 - `name: "web_search"`
637
638 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
639
640 Permission policy for tool execution.
641
642 - `BetaManagedAgentsAlwaysAllowPolicy object`
643
644 Tool calls are automatically approved without user confirmation.
645
646 - `BetaManagedAgentsAlwaysAskPolicy object`
647
648 Tool calls require user confirmation before execution.
649
650 - `BetaManagedAgentsAutoPolicy object`
651
652 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
653
654 - `allowed_domains: optional array of string`
655
656 - `blocked_domains: optional array of string`
657
658 - `user_location: optional BetaManagedAgentsUserLocation or null`
659
660 Approximate user location for search result localization.
661
662 - `type: "approximate"`
663
664 Location precision. Only "approximate" is supported.
665
666 - `city: optional string or null`
667
668 City name.
669
670 minLength: 1, maxLength: 255
671
672 - `country: optional string or null`
673
674 Two-letter ISO 3166-1 country code, uppercase.
675
676 - `region: optional string or null`
677
678 Region or state name.
679
680 minLength: 1, maxLength: 255
681
682 - `timezone: optional string or null`
683
684 IANA timezone identifier, e.g. "America/Los_Angeles".
685
686 minLength: 1, maxLength: 255
687
688 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
689
690 Resolved default configuration for agent tools.
361691
362692 - `enabled: boolean`
363693
364 - `name: "bash"`
365
366694 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
367695
368696 Permission policy for tool execution.
from line 699
371699
372700 Tool calls are automatically approved without user confirmation.
373701
374 - `type: "always_allow"`
375
376702 - `BetaManagedAgentsAlwaysAskPolicy object`
377703
378704 Tool calls require user confirmation before execution.
379705
380 - `type: "always_ask"`
381
382706 - `BetaManagedAgentsAutoPolicy object`
383707
384708 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
385709
386 - `type: "auto"`
387
388 - `BetaManagedAgentsEditToolConfig object`
389
390 Configuration for the edit tool.
391
392 - `type: "edit"`
710 - `BetaManagedAgentsMCPToolset object`
711
712 - `type: "mcp_toolset"`
713
714 - `configs: array of BetaManagedAgentsMCPToolConfig`
393715
394716 - `enabled: boolean`
395717
396 - `name: "edit"`
718 - `name: string`
397719
398720 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
399721
from line 733
411733
412734 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
413735
414 - `BetaManagedAgentsReadToolConfig object`
415
416 Configuration for the read tool.
417
418 - `type: "read"`
736 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
737
738 Resolved default configuration for all tools from an MCP server.
419739
420740 - `enabled: boolean`
421741
422 - `name: "read"`
423
424742 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
425743
426744 Permission policy for tool execution.
from line 755
437755
438756 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
439757
440 - `BetaManagedAgentsWriteToolConfig object`
441
442 Configuration for the write tool.
443
444 - `type: "write"`
445
446 - `enabled: boolean`
447
448 - `name: "write"`
449
450 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
451
452 Permission policy for tool execution.
453
454 - `BetaManagedAgentsAlwaysAllowPolicy object`
455
456 Tool calls are automatically approved without user confirmation.
457
458 - `BetaManagedAgentsAlwaysAskPolicy object`
459
460 Tool calls require user confirmation before execution.
461
462 - `BetaManagedAgentsAutoPolicy object`
463
464 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
465
466 - `BetaManagedAgentsGlobToolConfig object`
467
468 Configuration for the glob tool.
469
470 - `type: "glob"`
471
472 - `enabled: boolean`
473
474 - `name: "glob"`
475
476 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
477
478 Permission policy for tool execution.
479
480 - `BetaManagedAgentsAlwaysAllowPolicy object`
481
482 Tool calls are automatically approved without user confirmation.
483
484 - `BetaManagedAgentsAlwaysAskPolicy object`
485
486 Tool calls require user confirmation before execution.
487
488 - `BetaManagedAgentsAutoPolicy object`
489
490 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
491
492 - `BetaManagedAgentsGrepToolConfig object`
493
494 Configuration for the grep tool.
495
496 - `type: "grep"`
497
498 - `enabled: boolean`
499
500 - `name: "grep"`
501
502 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
503
504 Permission policy for tool execution.
505
506 - `BetaManagedAgentsAlwaysAllowPolicy object`
507
508 Tool calls are automatically approved without user confirmation.
509
510 - `BetaManagedAgentsAlwaysAskPolicy object`
511
512 Tool calls require user confirmation before execution.
513
514 - `BetaManagedAgentsAutoPolicy object`
515
516 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
517
518 - `BetaManagedAgentsWebFetchToolConfig object`
519
520 Configuration for the web_fetch tool.
521
522 - `type: "web_fetch"`
523
524 - `enabled: boolean`
525
526 - `name: "web_fetch"`
527
528 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
529
530 Permission policy for tool execution.
531
532 - `BetaManagedAgentsAlwaysAllowPolicy object`
533
534 Tool calls are automatically approved without user confirmation.
535
536 - `BetaManagedAgentsAlwaysAskPolicy object`
537
538 Tool calls require user confirmation before execution.
539
540 - `BetaManagedAgentsAutoPolicy object`
541
542 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
543
544 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
545
546 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
547
548 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
549
550 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
551
552 - `BetaManagedAgentsWebFetchURLSourceAll object`
553
554 Every URL from this source may be fetched. This is the default.
555
556 - `type: "all"`
557
558 - `BetaManagedAgentsWebFetchURLSourceNone object`
559
560 This source contributes no URLs that may be fetched.
561
562 - `type: "none"`
563
564 - `BetaManagedAgentsWebFetchURLSourceOnly object`
565
566 Only the named tools' results contribute URLs that may be fetched.
567
568 - `type: "only"`
569
570 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
571
572 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
573
574 - `type: "tool_reference"`
575
576 Must be "tool_reference".
577
578 - `name: string`
579
580 Name of the tool. Compared exactly, so upper and lower case letters are different.
581
582 minLength: 1, maxLength: 128
583
584 - `BetaManagedAgentsWebFetchURLSourceExcept object`
585
586 Every tool's results contribute URLs that may be fetched, except the named tools' results.
587
588 - `type: "except"`
589
590 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
591
592 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
593
594 - `type: "tool_reference"`
595
596 Must be "tool_reference".
597
598 - `name: string`
599
600 Name of the tool. Compared exactly, so upper and lower case letters are different.
601
602 minLength: 1, maxLength: 128
603
604 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
605
606 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
607
608 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
609
610 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
611
612 - `BetaManagedAgentsWebFetchURLSourceAll object`
613
614 Every URL from this source may be fetched. This is the default.
615
616 - `BetaManagedAgentsWebFetchURLSourceNone object`
617
618 This source contributes no URLs that may be fetched.
619
620 - `allowed_domains: optional array of string`
621
622 - `blocked_domains: optional array of string`
623
624 - `max_content_tokens: optional number or null`
625
626 format: int32
627
628 - `BetaManagedAgentsWebSearchToolConfig object`
629
630 Configuration for the web_search tool.
631
632 - `type: "web_search"`
633
634 - `enabled: boolean`
635
636 - `name: "web_search"`
637
638 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
639
640 Permission policy for tool execution.
641
642 - `BetaManagedAgentsAlwaysAllowPolicy object`
643
644 Tool calls are automatically approved without user confirmation.
645
646 - `BetaManagedAgentsAlwaysAskPolicy object`
647
648 Tool calls require user confirmation before execution.
649
650 - `BetaManagedAgentsAutoPolicy object`
651
652 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
653
654 - `allowed_domains: optional array of string`
655
656 - `blocked_domains: optional array of string`
657
658 - `user_location: optional BetaManagedAgentsUserLocation or null`
659
660 Approximate user location for search result localization.
661
662 - `type: "approximate"`
663
664 Location precision. Only "approximate" is supported.
665
666 - `city: optional string or null`
667
668 City name.
669
670 minLength: 1, maxLength: 255
671
672 - `country: optional string or null`
673
674 Two-letter ISO 3166-1 country code, uppercase.
675
676 - `region: optional string or null`
677
678 Region or state name.
679
680 minLength: 1, maxLength: 255
681
682 - `timezone: optional string or null`
683
684 IANA timezone identifier, e.g. "America/Los_Angeles".
685
686 minLength: 1, maxLength: 255
687
688 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
689
690 Resolved default configuration for agent tools.
691
692 - `enabled: boolean`
693
694 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
695
696 Permission policy for tool execution.
697
698 - `BetaManagedAgentsAlwaysAllowPolicy object`
699
700 Tool calls are automatically approved without user confirmation.
701
702 - `BetaManagedAgentsAlwaysAskPolicy object`
703
704 Tool calls require user confirmation before execution.
705
706 - `BetaManagedAgentsAutoPolicy object`
707
708 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
709
710 - `BetaManagedAgentsMCPToolset object`
711
712 - `type: "mcp_toolset"`
713
714 - `configs: array of BetaManagedAgentsMCPToolConfig`
715
716 - `enabled: boolean`
758 - `mcp_server_name: string`
759
760 - `BetaManagedAgentsCustomTool object`
761
762 A custom tool as returned in API responses.
763
764 - `type: "custom"`
765
766 - `description: string`
767
768 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
769
770 JSON Schema for custom tool input parameters.
771
772 - `type: "object"`
773
774 - `properties: optional map[unknown] or null`
775
776 - `required: optional array of string or null`
717777
718778 - `name: string`
719779
720 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
721
722 Permission policy for tool execution.
723
724 - `BetaManagedAgentsAlwaysAllowPolicy object`
725
726 Tool calls are automatically approved without user confirmation.
727
728 - `BetaManagedAgentsAlwaysAskPolicy object`
729
730 Tool calls require user confirmation before execution.
731
732 - `BetaManagedAgentsAutoPolicy object`
733
734 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
735
736 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
737
738 Resolved default configuration for all tools from an MCP server.
739
740 - `enabled: boolean`
741
742 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
743
744 Permission policy for tool execution.
745
746 - `BetaManagedAgentsAlwaysAllowPolicy object`
747
748 Tool calls are automatically approved without user confirmation.
749
750 - `BetaManagedAgentsAlwaysAskPolicy object`
751
752 Tool calls require user confirmation before execution.
753
754 - `BetaManagedAgentsAutoPolicy object`
755
756 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
757
758 - `mcp_server_name: string`
759
760 - `BetaManagedAgentsCustomTool object`
761
762 A custom tool as returned in API responses.
763
764 - `type: "custom"`
765
766 - `description: string`
767
768 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
769
770 JSON Schema for custom tool input parameters.
771
772 - `type: "object"`
773
774 - `properties: optional map[unknown] or null`
775
776 - `required: optional array of string or null`
780 - `version: number`
781
782 format: int32
783
784 - `BetaManagedAgentsAdvisor object`
785
786 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
787
788 - `type: "advisor"`
789
790 - `model: string`
791
792 The advisor model id.
793
794 - `BetaManagedAgentsSessionMultiagent20261001 object`
795
796 Resolved multiagent configuration with three members, as copied to the `session` at creation.
797
798 - `type: "multiagent_20261001"`
799
800 - `advisor: BetaManagedAgentsMultiagentAdvisor`
801
802 Whether the session's primary thread can consult an advisor model.
803
804 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
805
806 The session's primary thread can consult `model` mid-turn.
807
808 - `type: "enabled"`
809
810 - `model: string`
811
812 The advisor model id.
813
814 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
815
816 The agent has no advisor.
817
818 - `type: "disabled"`
819
820 - `subagents: BetaManagedAgentsSessionMultiagentSubagents`
821
822 Whether the agent can spawn session threads.
823
824 - `BetaManagedAgentsSessionMultiagentSubagentsEnabled object`
825
826 The agent can spawn session threads.
827
828 - `type: "enabled"`
829
830 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
831
832 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
833
834 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
835
836 The agent can define inline agents.
837
838 - `type: "enabled"`
839
840 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
841
842 The agent cannot define inline agents.
843
844 - `type: "disabled"`
845
846 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
847
848 Full `agent` definitions of the predefined agents, which are saved agents that this agent can spawn as session threads.
849
850 - `type: "agent"`
851
852 - `id: string`
853
854 - `description: string or null`
855
856 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
857
858 - `model: BetaManagedAgentsModelConfig`
859
860 Model identifier and configuration.
777861
778862 - `name: string`
779863
780 - `version: number`
781
782 format: int32
783
784 - `BetaManagedAgentsAdvisor object`
785
786 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
787
788 - `type: "advisor"`
789
790 - `model: string`
791
792 The advisor model id.
864 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
865
866 - `system: string or null`
867
868 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
869
870 - `version: number`
871
872 format: int32
873
874 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
875
876 The agent cannot spawn session threads.
877
878 - `type: "disabled"`
879
880 - `workflows: BetaManagedAgentsSessionMultiagentWorkflows`
881
882 Whether the agent can start workflow runs.
883
884 - `BetaManagedAgentsSessionMultiagentWorkflowsEnabled object`
885
886 The agent can start workflow runs.
887
888 - `type: "enabled"`
889
890 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
891
892 Whether a run's plan can define inline agents, which are not saved.
893
894 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
895
896 Full `agent` definitions of the predefined agents, which are saved agents that a run's plan can use.
897
898 - `type: "agent"`
899
900 - `id: string`
901
902 - `description: string or null`
903
904 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
905
906 - `model: BetaManagedAgentsModelConfig`
907
908 Model identifier and configuration.
909
910 - `name: string`
911
912 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
913
914 - `system: string or null`
915
916 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
917
918 - `version: number`
919
920 format: int32
921
922 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
923
924 The agent cannot start workflow runs.
925
926 - `type: "disabled"`
793927
794928 - `name: string`
795929
796930
api/beta/sessions/create Changed · +571 / -441 lines
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 237
237237
238238 - `"claude-haiku-4-5"`
239239
240 Fastest model with near-frontier intelligence
241
242240 - `"claude-haiku-4-5-20251001"`
243241
244 Fastest model with near-frontier intelligence
245
246242 - `"claude-opus-4-5"`
247243
248244 Powerful intelligence for long-running agents and coding
from line 1333
13371333
13381334 - `"claude-haiku-4-5"`
13391335
1340 Fastest model with near-frontier intelligence
1341
13421336 - `"claude-haiku-4-5-20251001"`
13431337
1344 Fastest model with near-frontier intelligence
1345
13461338 - `"claude-opus-4-5"`
13471339
13481340 Powerful intelligence for long-running agents and coding
from line 1403
14111403
14121404 - `"fast"`
14131405
1414 - `multiagent: BetaManagedAgentsSessionMultiagentCoordinator or null`
1406 - `multiagent: BetaManagedAgentsSessionMultiagent or null`
14151407
14161408 Resolved multiagent orchestration configuration. Null when the agent is single-threaded.
14171409
1418 - `type: "coordinator"`
1419
1420 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
1421
1422 Full `agent` definitions the coordinator may spawn as session threads.
1423
1424 - `BetaManagedAgentsSessionThreadAgent object`
1425
1426 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
1427
1428 - `type: "agent"`
1429
1430 - `id: string`
1431
1432 - `description: string or null`
1433
1434 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
1435
1436 - `type: "url"`
1410 - `BetaManagedAgentsSessionMultiagentCoordinator object`
1411
1412 Resolved coordinator topology with full agent definitions for each roster member.
1413
1414 - `type: "coordinator"`
1415
1416 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
1417
1418 Full `agent` definitions the coordinator may spawn as session threads.
1419
1420 - `BetaManagedAgentsSessionThreadAgent object`
1421
1422 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
1423
1424 - `type: "agent"`
1425
1426 - `id: string`
1427
1428 - `description: string or null`
1429
1430 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
1431
1432 - `type: "url"`
1433
1434 - `name: string`
1435
1436 - `url: string`
1437
1438 - `model: BetaManagedAgentsModelConfig`
1439
1440 Model identifier and configuration.
14371441
14381442 - `name: string`
14391443
1440 - `url: string`
1441
1442 - `model: BetaManagedAgentsModelConfig`
1443
1444 Model identifier and configuration.
1445
1446 - `name: string`
1447
1448 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
1449
1450 - `BetaManagedAgentsAnthropicSkill object`
1451
1452 A resolved Anthropic-managed skill.
1453
1454 - `type: "anthropic"`
1455
1456 - `skill_id: string`
1457
1458 - `version: string`
1459
1460 - `BetaManagedAgentsCustomSkill object`
1461
1462 A resolved user-created custom skill.
1463
1464 - `type: "custom"`
1465
1466 - `skill_id: string`
1467
1468 - `version: string`
1469
1470 - `system: string or null`
1471
1472 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
1473
1474 - `BetaManagedAgentsAgentToolset20260401 object`
1475
1476 - `type: "agent_toolset_20260401"`
1477
1478 - `configs: array of BetaManagedAgentsAgentToolConfig`
1479
1480 - `BetaManagedAgentsBashToolConfig object`
1481
1482 Configuration for the bash tool.
1483
1484 - `type: "bash"`
1444 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
1445
1446 - `BetaManagedAgentsAnthropicSkill object`
1447
1448 A resolved Anthropic-managed skill.
1449
1450 - `type: "anthropic"`
1451
1452 - `skill_id: string`
1453
1454 - `version: string`
1455
1456 - `BetaManagedAgentsCustomSkill object`
1457
1458 A resolved user-created custom skill.
1459
1460 - `type: "custom"`
1461
1462 - `skill_id: string`
1463
1464 - `version: string`
1465
1466 - `system: string or null`
1467
1468 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
1469
1470 - `BetaManagedAgentsAgentToolset20260401 object`
1471
1472 - `type: "agent_toolset_20260401"`
1473
1474 - `configs: array of BetaManagedAgentsAgentToolConfig`
1475
1476 - `BetaManagedAgentsBashToolConfig object`
1477
1478 Configuration for the bash tool.
1479
1480 - `type: "bash"`
1481
1482 - `enabled: boolean`
1483
1484 - `name: "bash"`
1485
1486 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1487
1488 Permission policy for tool execution.
1489
1490 - `BetaManagedAgentsAlwaysAllowPolicy object`
1491
1492 Tool calls are automatically approved without user confirmation.
1493
1494 - `type: "always_allow"`
1495
1496 - `BetaManagedAgentsAlwaysAskPolicy object`
1497
1498 Tool calls require user confirmation before execution.
1499
1500 - `type: "always_ask"`
1501
1502 - `BetaManagedAgentsAutoPolicy object`
1503
1504 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1505
1506 - `type: "auto"`
1507
1508 - `BetaManagedAgentsEditToolConfig object`
1509
1510 Configuration for the edit tool.
1511
1512 - `type: "edit"`
1513
1514 - `enabled: boolean`
1515
1516 - `name: "edit"`
1517
1518 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1519
1520 Permission policy for tool execution.
1521
1522 - `BetaManagedAgentsAlwaysAllowPolicy object`
1523
1524 Tool calls are automatically approved without user confirmation.
1525
1526 - `BetaManagedAgentsAlwaysAskPolicy object`
1527
1528 Tool calls require user confirmation before execution.
1529
1530 - `BetaManagedAgentsAutoPolicy object`
1531
1532 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1533
1534 - `BetaManagedAgentsReadToolConfig object`
1535
1536 Configuration for the read tool.
1537
1538 - `type: "read"`
1539
1540 - `enabled: boolean`
1541
1542 - `name: "read"`
1543
1544 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1545
1546 Permission policy for tool execution.
1547
1548 - `BetaManagedAgentsAlwaysAllowPolicy object`
1549
1550 Tool calls are automatically approved without user confirmation.
1551
1552 - `BetaManagedAgentsAlwaysAskPolicy object`
1553
1554 Tool calls require user confirmation before execution.
1555
1556 - `BetaManagedAgentsAutoPolicy object`
1557
1558 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1559
1560 - `BetaManagedAgentsWriteToolConfig object`
1561
1562 Configuration for the write tool.
1563
1564 - `type: "write"`
1565
1566 - `enabled: boolean`
1567
1568 - `name: "write"`
1569
1570 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1571
1572 Permission policy for tool execution.
1573
1574 - `BetaManagedAgentsAlwaysAllowPolicy object`
1575
1576 Tool calls are automatically approved without user confirmation.
1577
1578 - `BetaManagedAgentsAlwaysAskPolicy object`
1579
1580 Tool calls require user confirmation before execution.
1581
1582 - `BetaManagedAgentsAutoPolicy object`
1583
1584 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1585
1586 - `BetaManagedAgentsGlobToolConfig object`
1587
1588 Configuration for the glob tool.
1589
1590 - `type: "glob"`
1591
1592 - `enabled: boolean`
1593
1594 - `name: "glob"`
1595
1596 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1597
1598 Permission policy for tool execution.
1599
1600 - `BetaManagedAgentsAlwaysAllowPolicy object`
1601
1602 Tool calls are automatically approved without user confirmation.
1603
1604 - `BetaManagedAgentsAlwaysAskPolicy object`
1605
1606 Tool calls require user confirmation before execution.
1607
1608 - `BetaManagedAgentsAutoPolicy object`
1609
1610 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1611
1612 - `BetaManagedAgentsGrepToolConfig object`
1613
1614 Configuration for the grep tool.
1615
1616 - `type: "grep"`
1617
1618 - `enabled: boolean`
1619
1620 - `name: "grep"`
1621
1622 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1623
1624 Permission policy for tool execution.
1625
1626 - `BetaManagedAgentsAlwaysAllowPolicy object`
1627
1628 Tool calls are automatically approved without user confirmation.
1629
1630 - `BetaManagedAgentsAlwaysAskPolicy object`
1631
1632 Tool calls require user confirmation before execution.
1633
1634 - `BetaManagedAgentsAutoPolicy object`
1635
1636 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1637
1638 - `BetaManagedAgentsWebFetchToolConfig object`
1639
1640 Configuration for the web_fetch tool.
1641
1642 - `type: "web_fetch"`
1643
1644 - `enabled: boolean`
1645
1646 - `name: "web_fetch"`
1647
1648 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1649
1650 Permission policy for tool execution.
1651
1652 - `BetaManagedAgentsAlwaysAllowPolicy object`
1653
1654 Tool calls are automatically approved without user confirmation.
1655
1656 - `BetaManagedAgentsAlwaysAskPolicy object`
1657
1658 Tool calls require user confirmation before execution.
1659
1660 - `BetaManagedAgentsAutoPolicy object`
1661
1662 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1663
1664 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
1665
1666 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
1667
1668 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
1669
1670 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
1671
1672 - `BetaManagedAgentsWebFetchURLSourceAll object`
1673
1674 Every URL from this source may be fetched. This is the default.
1675
1676 - `type: "all"`
1677
1678 - `BetaManagedAgentsWebFetchURLSourceNone object`
1679
1680 This source contributes no URLs that may be fetched.
1681
1682 - `type: "none"`
1683
1684 - `BetaManagedAgentsWebFetchURLSourceOnly object`
1685
1686 Only the named tools' results contribute URLs that may be fetched.
1687
1688 - `type: "only"`
1689
1690 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
1691
1692 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
1693
1694 - `type: "tool_reference"`
1695
1696 Must be "tool_reference".
1697
1698 - `name: string`
1699
1700 Name of the tool. Compared exactly, so upper and lower case letters are different.
1701
1702 minLength: 1, maxLength: 128
1703
1704 - `BetaManagedAgentsWebFetchURLSourceExcept object`
1705
1706 Every tool's results contribute URLs that may be fetched, except the named tools' results.
1707
1708 - `type: "except"`
1709
1710 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
1711
1712 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
1713
1714 - `type: "tool_reference"`
1715
1716 Must be "tool_reference".
1717
1718 - `name: string`
1719
1720 Name of the tool. Compared exactly, so upper and lower case letters are different.
1721
1722 minLength: 1, maxLength: 128
1723
1724 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
1725
1726 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
1727
1728 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
1729
1730 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
1731
1732 - `BetaManagedAgentsWebFetchURLSourceAll object`
1733
1734 Every URL from this source may be fetched. This is the default.
1735
1736 - `BetaManagedAgentsWebFetchURLSourceNone object`
1737
1738 This source contributes no URLs that may be fetched.
1739
1740 - `allowed_domains: optional array of string`
1741
1742 - `blocked_domains: optional array of string`
1743
1744 - `max_content_tokens: optional number or null`
1745
1746 format: int32
1747
1748 - `BetaManagedAgentsWebSearchToolConfig object`
1749
1750 Configuration for the web_search tool.
1751
1752 - `type: "web_search"`
1753
1754 - `enabled: boolean`
1755
1756 - `name: "web_search"`
1757
1758 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1759
1760 Permission policy for tool execution.
1761
1762 - `BetaManagedAgentsAlwaysAllowPolicy object`
1763
1764 Tool calls are automatically approved without user confirmation.
1765
1766 - `BetaManagedAgentsAlwaysAskPolicy object`
1767
1768 Tool calls require user confirmation before execution.
1769
1770 - `BetaManagedAgentsAutoPolicy object`
1771
1772 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1773
1774 - `allowed_domains: optional array of string`
1775
1776 - `blocked_domains: optional array of string`
1777
1778 - `user_location: optional BetaManagedAgentsUserLocation or null`
1779
1780 Approximate user location for search result localization.
1781
1782 - `type: "approximate"`
1783
1784 Location precision. Only "approximate" is supported.
1785
1786 - `city: optional string or null`
1787
1788 City name.
1789
1790 minLength: 1, maxLength: 255
1791
1792 - `country: optional string or null`
1793
1794 Two-letter ISO 3166-1 country code, uppercase.
1795
1796 - `region: optional string or null`
1797
1798 Region or state name.
1799
1800 minLength: 1, maxLength: 255
1801
1802 - `timezone: optional string or null`
1803
1804 IANA timezone identifier, e.g. "America/Los_Angeles".
1805
1806 minLength: 1, maxLength: 255
1807
1808 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
1809
1810 Resolved default configuration for agent tools.
14851811
14861812 - `enabled: boolean`
14871813
1488 - `name: "bash"`
1489
14901814 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
14911815
14921816 Permission policy for tool execution.
from line 1819
14951819
14961820 Tool calls are automatically approved without user confirmation.
14971821
1498 - `type: "always_allow"`
1499
15001822 - `BetaManagedAgentsAlwaysAskPolicy object`
15011823
15021824 Tool calls require user confirmation before execution.
15031825
1504 - `type: "always_ask"`
1505
15061826 - `BetaManagedAgentsAutoPolicy object`
15071827
15081828 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
15091829
1510 - `type: "auto"`
1511
1512 - `BetaManagedAgentsEditToolConfig object`
1513
1514 Configuration for the edit tool.
1515
1516 - `type: "edit"`
1830 - `BetaManagedAgentsMCPToolset object`
1831
1832 - `type: "mcp_toolset"`
1833
1834 - `configs: array of BetaManagedAgentsMCPToolConfig`
15171835
15181836 - `enabled: boolean`
15191837
1520 - `name: "edit"`
1838 - `name: string`
15211839
15221840 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
15231841
from line 1853
15351853
15361854 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
15371855
1538 - `BetaManagedAgentsReadToolConfig object`
1539
1540 Configuration for the read tool.
1541
1542 - `type: "read"`
1856 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
1857
1858 Resolved default configuration for all tools from an MCP server.
15431859
15441860 - `enabled: boolean`
15451861
1546 - `name: "read"`
1547
15481862 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
15491863
15501864 Permission policy for tool execution.
from line 1875
15611875
15621876 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
15631877
1564 - `BetaManagedAgentsWriteToolConfig object`
1565
1566 Configuration for the write tool.
1567
1568 - `type: "write"`
1569
1570 - `enabled: boolean`
1571
1572 - `name: "write"`
1573
1574 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1575
1576 Permission policy for tool execution.
1577
1578 - `BetaManagedAgentsAlwaysAllowPolicy object`
1579
1580 Tool calls are automatically approved without user confirmation.
1581
1582 - `BetaManagedAgentsAlwaysAskPolicy object`
1583
1584 Tool calls require user confirmation before execution.
1585
1586 - `BetaManagedAgentsAutoPolicy object`
1587
1588 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1589
1590 - `BetaManagedAgentsGlobToolConfig object`
1591
1592 Configuration for the glob tool.
1593
1594 - `type: "glob"`
1595
1596 - `enabled: boolean`
1597
1598 - `name: "glob"`
1599
1600 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1601
1602 Permission policy for tool execution.
1603
1604 - `BetaManagedAgentsAlwaysAllowPolicy object`
1605
1606 Tool calls are automatically approved without user confirmation.
1607
1608 - `BetaManagedAgentsAlwaysAskPolicy object`
1609
1610 Tool calls require user confirmation before execution.
1611
1612 - `BetaManagedAgentsAutoPolicy object`
1613
1614 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1615
1616 - `BetaManagedAgentsGrepToolConfig object`
1617
1618 Configuration for the grep tool.
1619
1620 - `type: "grep"`
1621
1622 - `enabled: boolean`
1623
1624 - `name: "grep"`
1625
1626 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1627
1628 Permission policy for tool execution.
1629
1630 - `BetaManagedAgentsAlwaysAllowPolicy object`
1631
1632 Tool calls are automatically approved without user confirmation.
1633
1634 - `BetaManagedAgentsAlwaysAskPolicy object`
1635
1636 Tool calls require user confirmation before execution.
1637
1638 - `BetaManagedAgentsAutoPolicy object`
1639
1640 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1641
1642 - `BetaManagedAgentsWebFetchToolConfig object`
1643
1644 Configuration for the web_fetch tool.
1645
1646 - `type: "web_fetch"`
1647
1648 - `enabled: boolean`
1649
1650 - `name: "web_fetch"`
1651
1652 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1653
1654 Permission policy for tool execution.
1655
1656 - `BetaManagedAgentsAlwaysAllowPolicy object`
1657
1658 Tool calls are automatically approved without user confirmation.
1659
1660 - `BetaManagedAgentsAlwaysAskPolicy object`
1661
1662 Tool calls require user confirmation before execution.
1663
1664 - `BetaManagedAgentsAutoPolicy object`
1665
1666 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1667
1668 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
1669
1670 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
1671
1672 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
1673
1674 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
1675
1676 - `BetaManagedAgentsWebFetchURLSourceAll object`
1677
1678 Every URL from this source may be fetched. This is the default.
1679
1680 - `type: "all"`
1681
1682 - `BetaManagedAgentsWebFetchURLSourceNone object`
1683
1684 This source contributes no URLs that may be fetched.
1685
1686 - `type: "none"`
1687
1688 - `BetaManagedAgentsWebFetchURLSourceOnly object`
1689
1690 Only the named tools' results contribute URLs that may be fetched.
1691
1692 - `type: "only"`
1693
1694 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
1695
1696 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
1697
1698 - `type: "tool_reference"`
1699
1700 Must be "tool_reference".
1701
1702 - `name: string`
1703
1704 Name of the tool. Compared exactly, so upper and lower case letters are different.
1705
1706 minLength: 1, maxLength: 128
1707
1708 - `BetaManagedAgentsWebFetchURLSourceExcept object`
1709
1710 Every tool's results contribute URLs that may be fetched, except the named tools' results.
1711
1712 - `type: "except"`
1713
1714 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
1715
1716 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
1717
1718 - `type: "tool_reference"`
1719
1720 Must be "tool_reference".
1721
1722 - `name: string`
1723
1724 Name of the tool. Compared exactly, so upper and lower case letters are different.
1725
1726 minLength: 1, maxLength: 128
1727
1728 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
1729
1730 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
1731
1732 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
1733
1734 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
1735
1736 - `BetaManagedAgentsWebFetchURLSourceAll object`
1737
1738 Every URL from this source may be fetched. This is the default.
1739
1740 - `BetaManagedAgentsWebFetchURLSourceNone object`
1741
1742 This source contributes no URLs that may be fetched.
1743
1744 - `allowed_domains: optional array of string`
1745
1746 - `blocked_domains: optional array of string`
1747
1748 - `max_content_tokens: optional number or null`
1749
1750 format: int32
1751
1752 - `BetaManagedAgentsWebSearchToolConfig object`
1753
1754 Configuration for the web_search tool.
1755
1756 - `type: "web_search"`
1757
1758 - `enabled: boolean`
1759
1760 - `name: "web_search"`
1761
1762 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1763
1764 Permission policy for tool execution.
1765
1766 - `BetaManagedAgentsAlwaysAllowPolicy object`
1767
1768 Tool calls are automatically approved without user confirmation.
1769
1770 - `BetaManagedAgentsAlwaysAskPolicy object`
1771
1772 Tool calls require user confirmation before execution.
1773
1774 - `BetaManagedAgentsAutoPolicy object`
1775
1776 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1777
1778 - `allowed_domains: optional array of string`
1779
1780 - `blocked_domains: optional array of string`
1781
1782 - `user_location: optional BetaManagedAgentsUserLocation or null`
1783
1784 Approximate user location for search result localization.
1785
1786 - `type: "approximate"`
1787
1788 Location precision. Only "approximate" is supported.
1789
1790 - `city: optional string or null`
1791
1792 City name.
1793
1794 minLength: 1, maxLength: 255
1795
1796 - `country: optional string or null`
1797
1798 Two-letter ISO 3166-1 country code, uppercase.
1799
1800 - `region: optional string or null`
1801
1802 Region or state name.
1803
1804 minLength: 1, maxLength: 255
1805
1806 - `timezone: optional string or null`
1807
1808 IANA timezone identifier, e.g. "America/Los_Angeles".
1809
1810 minLength: 1, maxLength: 255
1811
1812 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
1813
1814 Resolved default configuration for agent tools.
1815
1816 - `enabled: boolean`
1817
1818 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1819
1820 Permission policy for tool execution.
1821
1822 - `BetaManagedAgentsAlwaysAllowPolicy object`
1823
1824 Tool calls are automatically approved without user confirmation.
1825
1826 - `BetaManagedAgentsAlwaysAskPolicy object`
1827
1828 Tool calls require user confirmation before execution.
1829
1830 - `BetaManagedAgentsAutoPolicy object`
1831
1832 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1833
1834 - `BetaManagedAgentsMCPToolset object`
1835
1836 - `type: "mcp_toolset"`
1837
1838 - `configs: array of BetaManagedAgentsMCPToolConfig`
1839
1840 - `enabled: boolean`
1878 - `mcp_server_name: string`
1879
1880 - `BetaManagedAgentsCustomTool object`
1881
1882 A custom tool as returned in API responses.
1883
1884 - `type: "custom"`
1885
1886 - `description: string`
1887
1888 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
1889
1890 JSON Schema for custom tool input parameters.
1891
1892 - `type: "object"`
1893
1894 - `properties: optional map[unknown] or null`
1895
1896 - `required: optional array of string or null`
18411897
18421898 - `name: string`
18431899
1844 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1845
1846 Permission policy for tool execution.
1847
1848 - `BetaManagedAgentsAlwaysAllowPolicy object`
1849
1850 Tool calls are automatically approved without user confirmation.
1851
1852 - `BetaManagedAgentsAlwaysAskPolicy object`
1853
1854 Tool calls require user confirmation before execution.
1855
1856 - `BetaManagedAgentsAutoPolicy object`
1857
1858 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1859
1860 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
1861
1862 Resolved default configuration for all tools from an MCP server.
1863
1864 - `enabled: boolean`
1865
1866 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
1867
1868 Permission policy for tool execution.
1869
1870 - `BetaManagedAgentsAlwaysAllowPolicy object`
1871
1872 Tool calls are automatically approved without user confirmation.
1873
1874 - `BetaManagedAgentsAlwaysAskPolicy object`
1875
1876 Tool calls require user confirmation before execution.
1877
1878 - `BetaManagedAgentsAutoPolicy object`
1879
1880 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
1881
1882 - `mcp_server_name: string`
1883
1884 - `BetaManagedAgentsCustomTool object`
1885
1886 A custom tool as returned in API responses.
1887
1888 - `type: "custom"`
1889
1890 - `description: string`
1891
1892 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
1893
1894 JSON Schema for custom tool input parameters.
1895
1896 - `type: "object"`
1897
1898 - `properties: optional map[unknown] or null`
1899
1900 - `required: optional array of string or null`
1900 - `version: number`
1901
1902 format: int32
1903
1904 - `BetaManagedAgentsAdvisor object`
1905
1906 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
1907
1908 - `type: "advisor"`
1909
1910 - `model: string`
1911
1912 The advisor model id.
1913
1914 - `BetaManagedAgentsSessionMultiagent20261001 object`
1915
1916 Resolved multiagent configuration with three members, as copied to the `session` at creation.
1917
1918 - `type: "multiagent_20261001"`
1919
1920 - `advisor: BetaManagedAgentsMultiagentAdvisor`
1921
1922 Whether the session's primary thread can consult an advisor model.
1923
1924 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
1925
1926 The session's primary thread can consult `model` mid-turn.
1927
1928 - `type: "enabled"`
1929
1930 - `model: string`
1931
1932 The advisor model id.
1933
1934 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
1935
1936 The agent has no advisor.
1937
1938 - `type: "disabled"`
1939
1940 - `subagents: BetaManagedAgentsSessionMultiagentSubagents`
1941
1942 Whether the agent can spawn session threads.
1943
1944 - `BetaManagedAgentsSessionMultiagentSubagentsEnabled object`
1945
1946 The agent can spawn session threads.
1947
1948 - `type: "enabled"`
1949
1950 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
1951
1952 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
1953
1954 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
1955
1956 The agent can define inline agents.
1957
1958 - `type: "enabled"`
1959
1960 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
1961
1962 The agent cannot define inline agents.
1963
1964 - `type: "disabled"`
1965
1966 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
1967
1968 Full `agent` definitions of the predefined agents, which are saved agents that this agent can spawn as session threads.
1969
1970 - `type: "agent"`
1971
1972 - `id: string`
1973
1974 - `description: string or null`
1975
1976 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
1977
1978 - `model: BetaManagedAgentsModelConfig`
1979
1980 Model identifier and configuration.
19011981
19021982 - `name: string`
19031983
1904 - `version: number`
1905
1906 format: int32
1907
1908 - `BetaManagedAgentsAdvisor object`
1909
1910 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
1911
1912 - `type: "advisor"`
1913
1914 - `model: string`
1915
1916 The advisor model id.
1984 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
1985
1986 - `system: string or null`
1987
1988 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
1989
1990 - `version: number`
1991
1992 format: int32
1993
1994 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
1995
1996 The agent cannot spawn session threads.
1997
1998 - `type: "disabled"`
1999
2000 - `workflows: BetaManagedAgentsSessionMultiagentWorkflows`
2001
2002 Whether the agent can start workflow runs.
2003
2004 - `BetaManagedAgentsSessionMultiagentWorkflowsEnabled object`
2005
2006 The agent can start workflow runs.
2007
2008 - `type: "enabled"`
2009
2010 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
2011
2012 Whether a run's plan can define inline agents, which are not saved.
2013
2014 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
2015
2016 Full `agent` definitions of the predefined agents, which are saved agents that a run's plan can use.
2017
2018 - `type: "agent"`
2019
2020 - `id: string`
2021
2022 - `description: string or null`
2023
2024 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
2025
2026 - `model: BetaManagedAgentsModelConfig`
2027
2028 Model identifier and configuration.
2029
2030 - `name: string`
2031
2032 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2033
2034 - `system: string or null`
2035
2036 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2037
2038 - `version: number`
2039
2040 format: int32
2041
2042 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
2043
2044 The agent cannot start workflow runs.
2045
2046 - `type: "disabled"`
19172047
19182048 - `name: string`
19192049
19202050
api/beta/sessions/events Changed · +7989 / -5771 lines
### Beta Managed Agents Max Workflow Runs Workflow Run Error ### Beta Managed Agents Program Workflow Run Error ### Beta Managed Agents Thread Limit Workflow Run Error ### Beta Managed Agents Timeout Workflow Run Error ### Beta Managed Agents Unknown Workflow Run Error ### Beta Managed Agents Workflow Run Created Event ### Beta Managed Agents Workflow Run Error ### Beta Managed Agents Workflow Run Error Event ### Beta Managed Agents Workflow Run Phase ### Beta Managed Agents Workflow Run Phase Ended Event ### Beta Managed Agents Workflow Run Phase Started Event ### Beta Managed Agents Workflow Run Result ### Beta Managed Agents Workflow Run Result Completed ### Beta Managed Agents Workflow Run Result Error ### Beta Managed Agents Workflow Run Result Stopped ### Beta Managed Agents Workflow Run Status Ended Event ### Beta Managed Agents Workflow Run Status Idle Event ### Beta Managed Agents Workflow Run Status Running Event
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
Nothing in the body moved in this read. What changed is above.
api/beta/sessions/events/list Changed · +847 / -439 lines
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 127
127127
128128 - `"session.usage"`
129129
130 - `"workflow_run.created"`
131
132 - `"workflow_run.status_running"`
133
134 - `"workflow_run.status_idle"`
135
136 - `"workflow_run.status_ended"`
137
138 - `"workflow_run.error"`
139
140 - `"workflow_run.phase_started"`
141
142 - `"workflow_run.phase_ended"`
143
130144## Headers
131145
132146- `"anthropic-version": optional string`
from line 499
485499
486500 - `custom_tool_use_id: string`
487501
488 The id of the `agent.custom_tool_use` event this result corresponds to, which can be found in the last `session.status_idle` [event's](https://platform.claude.com/docs/en/api/beta/sessions/events/list#beta_managed_agents_session_requires_action.event_ids) `stop_reason.event_ids` field.
502 The id of the `agent.custom_tool_use` event this result corresponds to. It is also listed in the last `session.status_idle` [event's](https://platform.claude.com/docs/en/api/beta/sessions/events/list#beta_managed_agents_session_requires_action.event_ids) `stop_reason.event_ids` field.
489503
490504 - `content: optional array of BetaManagedAgentsTextBlock or BetaManagedAgentsImageBlock or BetaManagedAgentsDocumentBlock or BetaManagedAgentsSearchResultBlock`
491505
from line 571
557571
558572 - `BetaManagedAgentsAgentCustomToolUseEvent object`
559573
560 Event emitted when the agent calls a custom tool. The session goes idle until the client sends a `user.custom_tool_result` event with the result.
574 Event emitted when the agent calls a custom tool. The session goes idle until the client sends a `user.custom_tool_result` event with the result. The client can send it as soon as this event arrives, without waiting for `session.status_idle`.
561575
562576 - `type: "agent.custom_tool_use"`
563577
from line 1513
14991513
15001514 Public `sthr_` ID of the newly created thread.
15011515
1516 - `workflow_run_id: string or null`
1517
1518 Identifier of the workflow run that created the thread, or `null` for any other thread.
1519
15021520 - `BetaManagedAgentsSpanOutcomeEvaluationStartEvent object`
15031521
15041522 Emitted when an outcome evaluation cycle begins.
from line 2017
19992017
20002018 - `"claude-haiku-4-5"`
20012019
2002 Fastest model with near-frontier intelligence
2003
20042020 - `"claude-haiku-4-5-20251001"`
20052021
2006 Fastest model with near-frontier intelligence
2007
20082022 - `"claude-opus-4-5"`
20092023
20102024 Powerful intelligence for long-running agents and coding
from line 2087
20732087
20742088 - `"fast"`
20752089
2076 - `multiagent: BetaManagedAgentsSessionMultiagentCoordinator or null`
2090 - `multiagent: BetaManagedAgentsSessionMultiagent or null`
20772091
20782092 Resolved multiagent orchestration configuration. Null when the agent is single-threaded.
20792093
2080 - `type: "coordinator"`
2081
2082 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
2083
2084 Full `agent` definitions the coordinator may spawn as session threads.
2085
2086 - `BetaManagedAgentsSessionThreadAgent object`
2087
2088 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
2089
2090 - `type: "agent"`
2091
2092 - `id: string`
2093
2094 - `description: string or null`
2095
2096 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
2097
2098 - `type: "url"`
2094 - `BetaManagedAgentsSessionMultiagentCoordinator object`
2095
2096 Resolved coordinator topology with full agent definitions for each roster member.
2097
2098 - `type: "coordinator"`
2099
2100 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
2101
2102 Full `agent` definitions the coordinator may spawn as session threads.
2103
2104 - `BetaManagedAgentsSessionThreadAgent object`
2105
2106 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
2107
2108 - `type: "agent"`
2109
2110 - `id: string`
2111
2112 - `description: string or null`
2113
2114 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
2115
2116 - `type: "url"`
2117
2118 - `name: string`
2119
2120 - `url: string`
2121
2122 - `model: BetaManagedAgentsModelConfig`
2123
2124 Model identifier and configuration.
20992125
21002126 - `name: string`
21012127
2102 - `url: string`
2103
2104 - `model: BetaManagedAgentsModelConfig`
2105
2106 Model identifier and configuration.
2107
2108 - `name: string`
2109
2110 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2111
2112 - `BetaManagedAgentsAnthropicSkill object`
2113
2114 A resolved Anthropic-managed skill.
2115
2116 - `type: "anthropic"`
2117
2118 - `skill_id: string`
2119
2120 - `version: string`
2121
2122 - `BetaManagedAgentsCustomSkill object`
2123
2124 A resolved user-created custom skill.
2125
2126 - `type: "custom"`
2127
2128 - `skill_id: string`
2129
2130 - `version: string`
2131
2132 - `system: string or null`
2133
2134 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2135
2136 - `BetaManagedAgentsAgentToolset20260401 object`
2137
2138 - `type: "agent_toolset_20260401"`
2139
2140 - `configs: array of BetaManagedAgentsAgentToolConfig`
2141
2142 - `BetaManagedAgentsBashToolConfig object`
2143
2144 Configuration for the bash tool.
2145
2146 - `type: "bash"`
2128 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2129
2130 - `BetaManagedAgentsAnthropicSkill object`
2131
2132 A resolved Anthropic-managed skill.
2133
2134 - `type: "anthropic"`
2135
2136 - `skill_id: string`
2137
2138 - `version: string`
2139
2140 - `BetaManagedAgentsCustomSkill object`
2141
2142 A resolved user-created custom skill.
2143
2144 - `type: "custom"`
2145
2146 - `skill_id: string`
2147
2148 - `version: string`
2149
2150 - `system: string or null`
2151
2152 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2153
2154 - `BetaManagedAgentsAgentToolset20260401 object`
2155
2156 - `type: "agent_toolset_20260401"`
2157
2158 - `configs: array of BetaManagedAgentsAgentToolConfig`
2159
2160 - `BetaManagedAgentsBashToolConfig object`
2161
2162 Configuration for the bash tool.
2163
2164 - `type: "bash"`
2165
2166 - `enabled: boolean`
2167
2168 - `name: "bash"`
2169
2170 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2171
2172 Permission policy for tool execution.
2173
2174 - `BetaManagedAgentsAlwaysAllowPolicy object`
2175
2176 Tool calls are automatically approved without user confirmation.
2177
2178 - `type: "always_allow"`
2179
2180 - `BetaManagedAgentsAlwaysAskPolicy object`
2181
2182 Tool calls require user confirmation before execution.
2183
2184 - `type: "always_ask"`
2185
2186 - `BetaManagedAgentsAutoPolicy object`
2187
2188 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2189
2190 - `type: "auto"`
2191
2192 - `BetaManagedAgentsEditToolConfig object`
2193
2194 Configuration for the edit tool.
2195
2196 - `type: "edit"`
2197
2198 - `enabled: boolean`
2199
2200 - `name: "edit"`
2201
2202 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2203
2204 Permission policy for tool execution.
2205
2206 - `BetaManagedAgentsAlwaysAllowPolicy object`
2207
2208 Tool calls are automatically approved without user confirmation.
2209
2210 - `BetaManagedAgentsAlwaysAskPolicy object`
2211
2212 Tool calls require user confirmation before execution.
2213
2214 - `BetaManagedAgentsAutoPolicy object`
2215
2216 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2217
2218 - `BetaManagedAgentsReadToolConfig object`
2219
2220 Configuration for the read tool.
2221
2222 - `type: "read"`
2223
2224 - `enabled: boolean`
2225
2226 - `name: "read"`
2227
2228 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2229
2230 Permission policy for tool execution.
2231
2232 - `BetaManagedAgentsAlwaysAllowPolicy object`
2233
2234 Tool calls are automatically approved without user confirmation.
2235
2236 - `BetaManagedAgentsAlwaysAskPolicy object`
2237
2238 Tool calls require user confirmation before execution.
2239
2240 - `BetaManagedAgentsAutoPolicy object`
2241
2242 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2243
2244 - `BetaManagedAgentsWriteToolConfig object`
2245
2246 Configuration for the write tool.
2247
2248 - `type: "write"`
2249
2250 - `enabled: boolean`
2251
2252 - `name: "write"`
2253
2254 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2255
2256 Permission policy for tool execution.
2257
2258 - `BetaManagedAgentsAlwaysAllowPolicy object`
2259
2260 Tool calls are automatically approved without user confirmation.
2261
2262 - `BetaManagedAgentsAlwaysAskPolicy object`
2263
2264 Tool calls require user confirmation before execution.
2265
2266 - `BetaManagedAgentsAutoPolicy object`
2267
2268 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2269
2270 - `BetaManagedAgentsGlobToolConfig object`
2271
2272 Configuration for the glob tool.
2273
2274 - `type: "glob"`
2275
2276 - `enabled: boolean`
2277
2278 - `name: "glob"`
2279
2280 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2281
2282 Permission policy for tool execution.
2283
2284 - `BetaManagedAgentsAlwaysAllowPolicy object`
2285
2286 Tool calls are automatically approved without user confirmation.
2287
2288 - `BetaManagedAgentsAlwaysAskPolicy object`
2289
2290 Tool calls require user confirmation before execution.
2291
2292 - `BetaManagedAgentsAutoPolicy object`
2293
2294 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2295
2296 - `BetaManagedAgentsGrepToolConfig object`
2297
2298 Configuration for the grep tool.
2299
2300 - `type: "grep"`
2301
2302 - `enabled: boolean`
2303
2304 - `name: "grep"`
2305
2306 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2307
2308 Permission policy for tool execution.
2309
2310 - `BetaManagedAgentsAlwaysAllowPolicy object`
2311
2312 Tool calls are automatically approved without user confirmation.
2313
2314 - `BetaManagedAgentsAlwaysAskPolicy object`
2315
2316 Tool calls require user confirmation before execution.
2317
2318 - `BetaManagedAgentsAutoPolicy object`
2319
2320 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2321
2322 - `BetaManagedAgentsWebFetchToolConfig object`
2323
2324 Configuration for the web_fetch tool.
2325
2326 - `type: "web_fetch"`
2327
2328 - `enabled: boolean`
2329
2330 - `name: "web_fetch"`
2331
2332 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2333
2334 Permission policy for tool execution.
2335
2336 - `BetaManagedAgentsAlwaysAllowPolicy object`
2337
2338 Tool calls are automatically approved without user confirmation.
2339
2340 - `BetaManagedAgentsAlwaysAskPolicy object`
2341
2342 Tool calls require user confirmation before execution.
2343
2344 - `BetaManagedAgentsAutoPolicy object`
2345
2346 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2347
2348 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
2349
2350 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
2351
2352 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2353
2354 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
2355
2356 - `BetaManagedAgentsWebFetchURLSourceAll object`
2357
2358 Every URL from this source may be fetched. This is the default.
2359
2360 - `type: "all"`
2361
2362 - `BetaManagedAgentsWebFetchURLSourceNone object`
2363
2364 This source contributes no URLs that may be fetched.
2365
2366 - `type: "none"`
2367
2368 - `BetaManagedAgentsWebFetchURLSourceOnly object`
2369
2370 Only the named tools' results contribute URLs that may be fetched.
2371
2372 - `type: "only"`
2373
2374 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2375
2376 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
2377
2378 - `type: "tool_reference"`
2379
2380 Must be "tool_reference".
2381
2382 - `name: string`
2383
2384 Name of the tool. Compared exactly, so upper and lower case letters are different.
2385
2386 minLength: 1, maxLength: 128
2387
2388 - `BetaManagedAgentsWebFetchURLSourceExcept object`
2389
2390 Every tool's results contribute URLs that may be fetched, except the named tools' results.
2391
2392 - `type: "except"`
2393
2394 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2395
2396 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
2397
2398 - `type: "tool_reference"`
2399
2400 Must be "tool_reference".
2401
2402 - `name: string`
2403
2404 Name of the tool. Compared exactly, so upper and lower case letters are different.
2405
2406 minLength: 1, maxLength: 128
2407
2408 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2409
2410 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
2411
2412 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
2413
2414 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
2415
2416 - `BetaManagedAgentsWebFetchURLSourceAll object`
2417
2418 Every URL from this source may be fetched. This is the default.
2419
2420 - `BetaManagedAgentsWebFetchURLSourceNone object`
2421
2422 This source contributes no URLs that may be fetched.
2423
2424 - `allowed_domains: optional array of string`
2425
2426 - `blocked_domains: optional array of string`
2427
2428 - `max_content_tokens: optional number or null`
2429
2430 format: int32
2431
2432 - `BetaManagedAgentsWebSearchToolConfig object`
2433
2434 Configuration for the web_search tool.
2435
2436 - `type: "web_search"`
2437
2438 - `enabled: boolean`
2439
2440 - `name: "web_search"`
2441
2442 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2443
2444 Permission policy for tool execution.
2445
2446 - `BetaManagedAgentsAlwaysAllowPolicy object`
2447
2448 Tool calls are automatically approved without user confirmation.
2449
2450 - `BetaManagedAgentsAlwaysAskPolicy object`
2451
2452 Tool calls require user confirmation before execution.
2453
2454 - `BetaManagedAgentsAutoPolicy object`
2455
2456 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2457
2458 - `allowed_domains: optional array of string`
2459
2460 - `blocked_domains: optional array of string`
2461
2462 - `user_location: optional BetaManagedAgentsUserLocation or null`
2463
2464 Approximate user location for search result localization.
2465
2466 - `type: "approximate"`
2467
2468 Location precision. Only "approximate" is supported.
2469
2470 - `city: optional string or null`
2471
2472 City name.
2473
2474 minLength: 1, maxLength: 255
2475
2476 - `country: optional string or null`
2477
2478 Two-letter ISO 3166-1 country code, uppercase.
2479
2480 - `region: optional string or null`
2481
2482 Region or state name.
2483
2484 minLength: 1, maxLength: 255
2485
2486 - `timezone: optional string or null`
2487
2488 IANA timezone identifier, e.g. "America/Los_Angeles".
2489
2490 minLength: 1, maxLength: 255
2491
2492 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
2493
2494 Resolved default configuration for agent tools.
21472495
21482496 - `enabled: boolean`
21492497
2150 - `name: "bash"`
2151
21522498 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
21532499
21542500 Permission policy for tool execution.
from line 2503
21572503
21582504 Tool calls are automatically approved without user confirmation.
21592505
2160 - `type: "always_allow"`
2161
21622506 - `BetaManagedAgentsAlwaysAskPolicy object`
21632507
21642508 Tool calls require user confirmation before execution.
21652509
2166 - `type: "always_ask"`
2167
21682510 - `BetaManagedAgentsAutoPolicy object`
21692511
21702512 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
21712513
2172 - `type: "auto"`
2173
2174 - `BetaManagedAgentsEditToolConfig object`
2175
2176 Configuration for the edit tool.
2177
2178 - `type: "edit"`
2514 - `BetaManagedAgentsMCPToolset object`
2515
2516 - `type: "mcp_toolset"`
2517
2518 - `configs: array of BetaManagedAgentsMCPToolConfig`
21792519
21802520 - `enabled: boolean`
21812521
2182 - `name: "edit"`
2522 - `name: string`
21832523
21842524 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
21852525
from line 2537
21972537
21982538 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
21992539
2200 - `BetaManagedAgentsReadToolConfig object`
2201
2202 Configuration for the read tool.
2203
2204 - `type: "read"`
2540 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
2541
2542 Resolved default configuration for all tools from an MCP server.
22052543
22062544 - `enabled: boolean`
22072545
2208 - `name: "read"`
2209
22102546 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
22112547
22122548 Permission policy for tool execution.
from line 2559
22232559
22242560 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
22252561
2226 - `BetaManagedAgentsWriteToolConfig object`
2227
2228 Configuration for the write tool.
2229
2230 - `type: "write"`
2231
2232 - `enabled: boolean`
2233
2234 - `name: "write"`
2235
2236 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2237
2238 Permission policy for tool execution.
2239
2240 - `BetaManagedAgentsAlwaysAllowPolicy object`
2241
2242 Tool calls are automatically approved without user confirmation.
2243
2244 - `BetaManagedAgentsAlwaysAskPolicy object`
2245
2246 Tool calls require user confirmation before execution.
2247
2248 - `BetaManagedAgentsAutoPolicy object`
2249
2250 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2251
2252 - `BetaManagedAgentsGlobToolConfig object`
2253
2254 Configuration for the glob tool.
2255
2256 - `type: "glob"`
2257
2258 - `enabled: boolean`
2259
2260 - `name: "glob"`
2261
2262 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2263
2264 Permission policy for tool execution.
2265
2266 - `BetaManagedAgentsAlwaysAllowPolicy object`
2267
2268 Tool calls are automatically approved without user confirmation.
2269
2270 - `BetaManagedAgentsAlwaysAskPolicy object`
2271
2272 Tool calls require user confirmation before execution.
2273
2274 - `BetaManagedAgentsAutoPolicy object`
2275
2276 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2277
2278 - `BetaManagedAgentsGrepToolConfig object`
2279
2280 Configuration for the grep tool.
2281
2282 - `type: "grep"`
2283
2284 - `enabled: boolean`
2285
2286 - `name: "grep"`
2287
2288 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2289
2290 Permission policy for tool execution.
2291
2292 - `BetaManagedAgentsAlwaysAllowPolicy object`
2293
2294 Tool calls are automatically approved without user confirmation.
2295
2296 - `BetaManagedAgentsAlwaysAskPolicy object`
2297
2298 Tool calls require user confirmation before execution.
2299
2300 - `BetaManagedAgentsAutoPolicy object`
2301
2302 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2303
2304 - `BetaManagedAgentsWebFetchToolConfig object`
2305
2306 Configuration for the web_fetch tool.
2307
2308 - `type: "web_fetch"`
2309
2310 - `enabled: boolean`
2311
2312 - `name: "web_fetch"`
2313
2314 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2315
2316 Permission policy for tool execution.
2317
2318 - `BetaManagedAgentsAlwaysAllowPolicy object`
2319
2320 Tool calls are automatically approved without user confirmation.
2321
2322 - `BetaManagedAgentsAlwaysAskPolicy object`
2323
2324 Tool calls require user confirmation before execution.
2325
2326 - `BetaManagedAgentsAutoPolicy object`
2327
2328 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2329
2330 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
2331
2332 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
2333
2334 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2335
2336 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
2337
2338 - `BetaManagedAgentsWebFetchURLSourceAll object`
2339
2340 Every URL from this source may be fetched. This is the default.
2341
2342 - `type: "all"`
2343
2344 - `BetaManagedAgentsWebFetchURLSourceNone object`
2345
2346 This source contributes no URLs that may be fetched.
2347
2348 - `type: "none"`
2349
2350 - `BetaManagedAgentsWebFetchURLSourceOnly object`
2351
2352 Only the named tools' results contribute URLs that may be fetched.
2353
2354 - `type: "only"`
2355
2356 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2357
2358 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
2359
2360 - `type: "tool_reference"`
2361
2362 Must be "tool_reference".
2363
2364 - `name: string`
2365
2366 Name of the tool. Compared exactly, so upper and lower case letters are different.
2367
2368 minLength: 1, maxLength: 128
2369
2370 - `BetaManagedAgentsWebFetchURLSourceExcept object`
2371
2372 Every tool's results contribute URLs that may be fetched, except the named tools' results.
2373
2374 - `type: "except"`
2375
2376 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2377
2378 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
2379
2380 - `type: "tool_reference"`
2381
2382 Must be "tool_reference".
2383
2384 - `name: string`
2385
2386 Name of the tool. Compared exactly, so upper and lower case letters are different.
2387
2388 minLength: 1, maxLength: 128
2389
2390 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2391
2392 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
2393
2394 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
2395
2396 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
2397
2398 - `BetaManagedAgentsWebFetchURLSourceAll object`
2399
2400 Every URL from this source may be fetched. This is the default.
2401
2402 - `BetaManagedAgentsWebFetchURLSourceNone object`
2403
2404 This source contributes no URLs that may be fetched.
2405
2406 - `allowed_domains: optional array of string`
2407
2408 - `blocked_domains: optional array of string`
2409
2410 - `max_content_tokens: optional number or null`
2411
2412 format: int32
2413
2414 - `BetaManagedAgentsWebSearchToolConfig object`
2415
2416 Configuration for the web_search tool.
2417
2418 - `type: "web_search"`
2419
2420 - `enabled: boolean`
2421
2422 - `name: "web_search"`
2423
2424 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2425
2426 Permission policy for tool execution.
2427
2428 - `BetaManagedAgentsAlwaysAllowPolicy object`
2429
2430 Tool calls are automatically approved without user confirmation.
2431
2432 - `BetaManagedAgentsAlwaysAskPolicy object`
2433
2434 Tool calls require user confirmation before execution.
2435
2436 - `BetaManagedAgentsAutoPolicy object`
2437
2438 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2439
2440 - `allowed_domains: optional array of string`
2441
2442 - `blocked_domains: optional array of string`
2443
2444 - `user_location: optional BetaManagedAgentsUserLocation or null`
2445
2446 Approximate user location for search result localization.
2447
2448 - `type: "approximate"`
2449
2450 Location precision. Only "approximate" is supported.
2451
2452 - `city: optional string or null`
2453
2454 City name.
2455
2456 minLength: 1, maxLength: 255
2457
2458 - `country: optional string or null`
2459
2460 Two-letter ISO 3166-1 country code, uppercase.
2461
2462 - `region: optional string or null`
2463
2464 Region or state name.
2465
2466 minLength: 1, maxLength: 255
2467
2468 - `timezone: optional string or null`
2469
2470 IANA timezone identifier, e.g. "America/Los_Angeles".
2471
2472 minLength: 1, maxLength: 255
2473
2474 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
2475
2476 Resolved default configuration for agent tools.
2477
2478 - `enabled: boolean`
2479
2480 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2481
2482 Permission policy for tool execution.
2483
2484 - `BetaManagedAgentsAlwaysAllowPolicy object`
2485
2486 Tool calls are automatically approved without user confirmation.
2487
2488 - `BetaManagedAgentsAlwaysAskPolicy object`
2489
2490 Tool calls require user confirmation before execution.
2491
2492 - `BetaManagedAgentsAutoPolicy object`
2493
2494 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2495
2496 - `BetaManagedAgentsMCPToolset object`
2497
2498 - `type: "mcp_toolset"`
2499
2500 - `configs: array of BetaManagedAgentsMCPToolConfig`
2501
2502 - `enabled: boolean`
2562 - `mcp_server_name: string`
2563
2564 - `BetaManagedAgentsCustomTool object`
2565
2566 A custom tool as returned in API responses.
2567
2568 - `type: "custom"`
2569
2570 - `description: string`
2571
2572 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
2573
2574 JSON Schema for custom tool input parameters.
2575
2576 - `type: "object"`
2577
2578 - `properties: optional map[unknown] or null`
2579
2580 - `required: optional array of string or null`
25032581
25042582 - `name: string`
25052583
2506 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2507
2508 Permission policy for tool execution.
2509
2510 - `BetaManagedAgentsAlwaysAllowPolicy object`
2511
2512 Tool calls are automatically approved without user confirmation.
2513
2514 - `BetaManagedAgentsAlwaysAskPolicy object`
2515
2516 Tool calls require user confirmation before execution.
2517
2518 - `BetaManagedAgentsAutoPolicy object`
2519
2520 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2521
2522 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
2523
2524 Resolved default configuration for all tools from an MCP server.
2525
2526 - `enabled: boolean`
2527
2528 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2529
2530 Permission policy for tool execution.
2531
2532 - `BetaManagedAgentsAlwaysAllowPolicy object`
2533
2534 Tool calls are automatically approved without user confirmation.
2535
2536 - `BetaManagedAgentsAlwaysAskPolicy object`
2537
2538 Tool calls require user confirmation before execution.
2539
2540 - `BetaManagedAgentsAutoPolicy object`
2541
2542 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2543
2544 - `mcp_server_name: string`
2545
2546 - `BetaManagedAgentsCustomTool object`
2547
2548 A custom tool as returned in API responses.
2549
2550 - `type: "custom"`
2551
2552 - `description: string`
2553
2554 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
2555
2556 JSON Schema for custom tool input parameters.
2557
2558 - `type: "object"`
2559
2560 - `properties: optional map[unknown] or null`
2561
2562 - `required: optional array of string or null`
2584 - `version: number`
2585
2586 format: int32
2587
2588 - `BetaManagedAgentsAdvisor object`
2589
2590 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
2591
2592 - `type: "advisor"`
2593
2594 - `model: string`
2595
2596 The advisor model id.
2597
2598 - `BetaManagedAgentsSessionMultiagent20261001 object`
2599
2600 Resolved multiagent configuration with three members, as copied to the `session` at creation.
2601
2602 - `type: "multiagent_20261001"`
2603
2604 - `advisor: BetaManagedAgentsMultiagentAdvisor`
2605
2606 Whether the session's primary thread can consult an advisor model.
2607
2608 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
2609
2610 The session's primary thread can consult `model` mid-turn.
2611
2612 - `type: "enabled"`
2613
2614 - `model: string`
2615
2616 The advisor model id.
2617
2618 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
2619
2620 The agent has no advisor.
2621
2622 - `type: "disabled"`
2623
2624 - `subagents: BetaManagedAgentsSessionMultiagentSubagents`
2625
2626 Whether the agent can spawn session threads.
2627
2628 - `BetaManagedAgentsSessionMultiagentSubagentsEnabled object`
2629
2630 The agent can spawn session threads.
2631
2632 - `type: "enabled"`
2633
2634 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
2635
2636 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
2637
2638 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
2639
2640 The agent can define inline agents.
2641
2642 - `type: "enabled"`
2643
2644 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
2645
2646 The agent cannot define inline agents.
2647
2648 - `type: "disabled"`
2649
2650 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
2651
2652 Full `agent` definitions of the predefined agents, which are saved agents that this agent can spawn as session threads.
2653
2654 - `type: "agent"`
2655
2656 - `id: string`
2657
2658 - `description: string or null`
2659
2660 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
2661
2662 - `model: BetaManagedAgentsModelConfig`
2663
2664 Model identifier and configuration.
25632665
25642666 - `name: string`
25652667
2566 - `version: number`
2567
2568 format: int32
2569
2570 - `BetaManagedAgentsAdvisor object`
2571
2572 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
2573
2574 - `type: "advisor"`
2575
2576 - `model: string`
2577
2578 The advisor model id.
2668 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2669
2670 - `system: string or null`
2671
2672 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2673
2674 - `version: number`
2675
2676 format: int32
2677
2678 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
2679
2680 The agent cannot spawn session threads.
2681
2682 - `type: "disabled"`
2683
2684 - `workflows: BetaManagedAgentsSessionMultiagentWorkflows`
2685
2686 Whether the agent can start workflow runs.
2687
2688 - `BetaManagedAgentsSessionMultiagentWorkflowsEnabled object`
2689
2690 The agent can start workflow runs.
2691
2692 - `type: "enabled"`
2693
2694 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
2695
2696 Whether a run's plan can define inline agents, which are not saved.
2697
2698 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
2699
2700 Full `agent` definitions of the predefined agents, which are saved agents that a run's plan can use.
2701
2702 - `type: "agent"`
2703
2704 - `id: string`
2705
2706 - `description: string or null`
2707
2708 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
2709
2710 - `model: BetaManagedAgentsModelConfig`
2711
2712 Model identifier and configuration.
2713
2714 - `name: string`
2715
2716 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2717
2718 - `system: string or null`
2719
2720 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2721
2722 - `version: number`
2723
2724 format: int32
2725
2726 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
2727
2728 The agent cannot start workflow runs.
2729
2730 - `type: "disabled"`
25792731
25802732 - `name: string`
25812733
from line 2894
27422894 - `budget: optional BetaManagedAgentsBudgetLimit or null`
27432895
27442896 The session's configured budget at the snapshot time, or null when the session has no budget.
2897
2898 - `BetaManagedAgentsWorkflowRunCreatedEvent object`
2899
2900 A workflow run was created. A workflow run is background work that the session's agent starts. Emitted once per run, before the run's other `workflow_run.*` events.
2901
2902 - `type: "workflow_run.created"`
2903
2904 - `id: string`
2905
2906 Unique identifier for this event.
2907
2908 - `description: string or null`
2909
2910 Description that the agent gave the run, passed on as written, or `null` if it gave none.
2911
2912 - `name: string`
2913
2914 Name that the agent gave the run, passed on as written, or a name that the server assigned.
2915
2916 - `phases: array of BetaManagedAgentsWorkflowRunPhase`
2917
2918 The phases that the run's plan declares, in the plan's order. Can be empty.
2919
2920 - `id: string`
2921
2922 Unique identifier for the phase.
2923
2924 - `description: string or null`
2925
2926 Description that the agent gave the phase, passed on as written, or `null` if it gave none.
2927
2928 - `name: string`
2929
2930 Name that the agent gave the phase, passed on as written.
2931
2932 - `processed_at: string`
2933
2934 Timestamp when this event was processed.
2935
2936 format: date-time
2937
2938 - `workflow_run_id: string`
2939
2940 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
2941
2942 - `BetaManagedAgentsWorkflowRunStatusEndedEvent object`
2943
2944 A workflow run ended. Emitted once per run, as the last of the run's `workflow_run.*` events.
2945
2946 - `type: "workflow_run.status_ended"`
2947
2948 - `id: string`
2949
2950 Unique identifier for this event.
2951
2952 - `processed_at: string`
2953
2954 Timestamp when this event was processed.
2955
2956 format: date-time
2957
2958 - `result: BetaManagedAgentsWorkflowRunResult`
2959
2960 How the run ended.
2961
2962 - `BetaManagedAgentsWorkflowRunResultCompleted object`
2963
2964 The run's plan, a program that the agent wrote, finished. This does not say whether the work succeeded.
2965
2966 - `type: "completed"`
2967
2968 - `BetaManagedAgentsWorkflowRunResultError object`
2969
2970 The run failed or reached its time limit.
2971
2972 - `type: "error"`
2973
2974 - `error: BetaManagedAgentsWorkflowRunError`
2975
2976 Why the run did not finish.
2977
2978 - `BetaManagedAgentsTimeoutWorkflowRunError object`
2979
2980 The run reached its time limit.
2981
2982 - `type: "timeout_error"`
2983
2984 - `message: string`
2985
2986 Short explanation written by the server. It never contains content from the run or its agents.
2987
2988 - `BetaManagedAgentsProgramWorkflowRunError object`
2989
2990 The plan, a program that the agent wrote, failed, or the server refused it.
2991
2992 - `type: "program_error"`
2993
2994 - `message: string`
2995
2996 Short explanation written by the server. It never contains content from the run or its agents.
2997
2998 - `BetaManagedAgentsUnknownWorkflowRunError object`
2999
3000 A failure that has no type of its own.
3001
3002 - `type: "unknown_error"`
3003
3004 - `message: string`
3005
3006 Short explanation written by the server. It never contains content from the run or its agents.
3007
3008 - `BetaManagedAgentsThreadLimitWorkflowRunError object`
3009
3010 The run exceeded the limit on the number of threads that a run can create.
3011
3012 - `type: "thread_limit_error"`
3013
3014 - `message: string`
3015
3016 Short explanation written by the server. It never contains content from the run or its agents.
3017
3018 - `BetaManagedAgentsMaxWorkflowRunsWorkflowRunError object`
3019
3020 No run was created, because the session was at its limit of open workflow runs, which are runs that have not ended. Only `workflow_run.error` carries this type.
3021
3022 - `type: "max_workflow_runs_error"`
3023
3024 - `message: string`
3025
3026 Short explanation written by the server. It never contains content from the run or its agents.
3027
3028 - `BetaManagedAgentsWorkflowRunResultStopped object`
3029
3030 The agent stopped the run.
3031
3032 - `type: "stopped"`
3033
3034 - `workflow_run_id: string`
3035
3036 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3037
3038 - `BetaManagedAgentsWorkflowRunPhaseStartedEvent object`
3039
3040 A workflow run's plan entered a phase.
3041
3042 - `type: "workflow_run.phase_started"`
3043
3044 - `id: string`
3045
3046 Unique identifier for this event.
3047
3048 - `processed_at: string`
3049
3050 Timestamp when this event was processed.
3051
3052 format: date-time
3053
3054 - `workflow_run_id: string`
3055
3056 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3057
3058 - `workflow_run_phase_id: string`
3059
3060 Identifier of the phase, as in `phases` on the run's `workflow_run.created` event.
3061
3062 - `BetaManagedAgentsWorkflowRunPhaseEndedEvent object`
3063
3064 A workflow run's plan left a phase, or the run's end closed it. Emitted once for every `workflow_run.phase_started` event, before the run's `workflow_run.status_ended` event. The event does not say whether the plan finished the phase's work, or why it left.
3065
3066 - `type: "workflow_run.phase_ended"`
3067
3068 - `id: string`
3069
3070 Unique identifier for this event.
3071
3072 - `phase_started_id: string`
3073
3074 Identifier of the `workflow_run.phase_started` event that opened the phase.
3075
3076 - `processed_at: string`
3077
3078 Timestamp when this event was processed.
3079
3080 format: date-time
3081
3082 - `workflow_run_id: string`
3083
3084 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3085
3086 - `workflow_run_phase_id: string`
3087
3088 Identifier of the phase, as in `phases` on the run's `workflow_run.created` event.
3089
3090 - `BetaManagedAgentsWorkflowRunStatusRunningEvent object`
3091
3092 A workflow run is running. Emitted when the run starts to execute, and each time it resumes after being idle. A run that starts idle emits `workflow_run.status_idle` first.
3093
3094 - `type: "workflow_run.status_running"`
3095
3096 - `id: string`
3097
3098 Unique identifier for this event.
3099
3100 - `processed_at: string`
3101
3102 Timestamp when this event was processed.
3103
3104 format: date-time
3105
3106 - `workflow_run_id: string`
3107
3108 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3109
3110 - `BetaManagedAgentsWorkflowRunStatusIdleEvent object`
3111
3112 A workflow run is idle. Emitted each time the run goes idle, whatever the cause. If the run ends while idle, no `workflow_run.status_running` comes between this event and its `workflow_run.status_ended`.
3113
3114 - `type: "workflow_run.status_idle"`
3115
3116 - `id: string`
3117
3118 Unique identifier for this event.
3119
3120 - `processed_at: string`
3121
3122 Timestamp when this event was processed.
3123
3124 format: date-time
3125
3126 - `workflow_run_id: string`
3127
3128 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3129
3130 - `BetaManagedAgentsWorkflowRunErrorEvent object`
3131
3132 A workflow run met an error, or an error kept a run from being created. A run that ends with a `result.type` of `error` emits this event before its `workflow_run.status_ended`, with the same `error`.
3133
3134 - `type: "workflow_run.error"`
3135
3136 - `id: string`
3137
3138 Unique identifier for this event.
3139
3140 - `error: BetaManagedAgentsWorkflowRunError`
3141
3142 Why the run did not finish, or was not created.
3143
3144 - `processed_at: string`
3145
3146 Timestamp when this event was processed.
3147
3148 format: date-time
3149
3150 - `workflow_run_id: string or null`
3151
3152 Identifier of the run that met the error, or `null` when the error kept a run from being created.
27453153
27463154- `next_page: optional string or null`
27473155
27483156
api/beta/sessions/events/stream Changed · +834 / -440 lines
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 141
141141
142142## Returns
143143
144- `BetaManagedAgentsStreamSessionEvents = BetaManagedAgentsUserMessageEvent or BetaManagedAgentsUserInterruptEvent or BetaManagedAgentsUserToolConfirmationEvent or 34 more`
144- `BetaManagedAgentsStreamSessionEvents = BetaManagedAgentsUserMessageEvent or BetaManagedAgentsUserInterruptEvent or BetaManagedAgentsUserToolConfirmationEvent or 41 more`
145145
146146 Server-sent event in the session stream.
147147
from line 381
381381
382382 - `custom_tool_use_id: string`
383383
384 The id of the `agent.custom_tool_use` event this result corresponds to, which can be found in the last `session.status_idle` [event's](https://platform.claude.com/docs/en/api/beta/sessions/events/list#beta_managed_agents_session_requires_action.event_ids) `stop_reason.event_ids` field.
384 The id of the `agent.custom_tool_use` event this result corresponds to. It is also listed in the last `session.status_idle` [event's](https://platform.claude.com/docs/en/api/beta/sessions/events/list#beta_managed_agents_session_requires_action.event_ids) `stop_reason.event_ids` field.
385385
386386 - `content: optional array of BetaManagedAgentsTextBlock or BetaManagedAgentsImageBlock or BetaManagedAgentsDocumentBlock or BetaManagedAgentsSearchResultBlock`
387387
from line 453
453453
454454 - `BetaManagedAgentsAgentCustomToolUseEvent object`
455455
456 Event emitted when the agent calls a custom tool. The session goes idle until the client sends a `user.custom_tool_result` event with the result.
456 Event emitted when the agent calls a custom tool. The session goes idle until the client sends a `user.custom_tool_result` event with the result. The client can send it as soon as this event arrives, without waiting for `session.status_idle`.
457457
458458 - `type: "agent.custom_tool_use"`
459459
from line 1395
13951395
13961396 Public `sthr_` ID of the newly created thread.
13971397
1398 - `workflow_run_id: string or null`
1399
1400 Identifier of the workflow run that created the thread, or `null` for any other thread.
1401
13981402 - `BetaManagedAgentsSpanOutcomeEvaluationStartEvent object`
13991403
14001404 Emitted when an outcome evaluation cycle begins.
from line 1899
18951899
18961900 - `"claude-haiku-4-5"`
18971901
1898 Fastest model with near-frontier intelligence
1899
19001902 - `"claude-haiku-4-5-20251001"`
19011903
1902 Fastest model with near-frontier intelligence
1903
19041904 - `"claude-opus-4-5"`
19051905
19061906 Powerful intelligence for long-running agents and coding
from line 1969
19691969
19701970 - `"fast"`
19711971
1972 - `multiagent: BetaManagedAgentsSessionMultiagentCoordinator or null`
1972 - `multiagent: BetaManagedAgentsSessionMultiagent or null`
19731973
19741974 Resolved multiagent orchestration configuration. Null when the agent is single-threaded.
19751975
1976 - `type: "coordinator"`
1977
1978 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
1979
1980 Full `agent` definitions the coordinator may spawn as session threads.
1981
1982 - `BetaManagedAgentsSessionThreadAgent object`
1983
1984 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
1985
1986 - `type: "agent"`
1987
1988 - `id: string`
1989
1990 - `description: string or null`
1991
1992 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
1993
1994 - `type: "url"`
1976 - `BetaManagedAgentsSessionMultiagentCoordinator object`
1977
1978 Resolved coordinator topology with full agent definitions for each roster member.
1979
1980 - `type: "coordinator"`
1981
1982 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
1983
1984 Full `agent` definitions the coordinator may spawn as session threads.
1985
1986 - `BetaManagedAgentsSessionThreadAgent object`
1987
1988 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
1989
1990 - `type: "agent"`
1991
1992 - `id: string`
1993
1994 - `description: string or null`
1995
1996 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
1997
1998 - `type: "url"`
1999
2000 - `name: string`
2001
2002 - `url: string`
2003
2004 - `model: BetaManagedAgentsModelConfig`
2005
2006 Model identifier and configuration.
19952007
19962008 - `name: string`
19972009
1998 - `url: string`
1999
2000 - `model: BetaManagedAgentsModelConfig`
2001
2002 Model identifier and configuration.
2003
2004 - `name: string`
2005
2006 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2007
2008 - `BetaManagedAgentsAnthropicSkill object`
2009
2010 A resolved Anthropic-managed skill.
2011
2012 - `type: "anthropic"`
2013
2014 - `skill_id: string`
2015
2016 - `version: string`
2017
2018 - `BetaManagedAgentsCustomSkill object`
2019
2020 A resolved user-created custom skill.
2021
2022 - `type: "custom"`
2023
2024 - `skill_id: string`
2025
2026 - `version: string`
2027
2028 - `system: string or null`
2029
2030 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2031
2032 - `BetaManagedAgentsAgentToolset20260401 object`
2033
2034 - `type: "agent_toolset_20260401"`
2035
2036 - `configs: array of BetaManagedAgentsAgentToolConfig`
2037
2038 - `BetaManagedAgentsBashToolConfig object`
2039
2040 Configuration for the bash tool.
2041
2042 - `type: "bash"`
2010 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2011
2012 - `BetaManagedAgentsAnthropicSkill object`
2013
2014 A resolved Anthropic-managed skill.
2015
2016 - `type: "anthropic"`
2017
2018 - `skill_id: string`
2019
2020 - `version: string`
2021
2022 - `BetaManagedAgentsCustomSkill object`
2023
2024 A resolved user-created custom skill.
2025
2026 - `type: "custom"`
2027
2028 - `skill_id: string`
2029
2030 - `version: string`
2031
2032 - `system: string or null`
2033
2034 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2035
2036 - `BetaManagedAgentsAgentToolset20260401 object`
2037
2038 - `type: "agent_toolset_20260401"`
2039
2040 - `configs: array of BetaManagedAgentsAgentToolConfig`
2041
2042 - `BetaManagedAgentsBashToolConfig object`
2043
2044 Configuration for the bash tool.
2045
2046 - `type: "bash"`
2047
2048 - `enabled: boolean`
2049
2050 - `name: "bash"`
2051
2052 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2053
2054 Permission policy for tool execution.
2055
2056 - `BetaManagedAgentsAlwaysAllowPolicy object`
2057
2058 Tool calls are automatically approved without user confirmation.
2059
2060 - `type: "always_allow"`
2061
2062 - `BetaManagedAgentsAlwaysAskPolicy object`
2063
2064 Tool calls require user confirmation before execution.
2065
2066 - `type: "always_ask"`
2067
2068 - `BetaManagedAgentsAutoPolicy object`
2069
2070 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2071
2072 - `type: "auto"`
2073
2074 - `BetaManagedAgentsEditToolConfig object`
2075
2076 Configuration for the edit tool.
2077
2078 - `type: "edit"`
2079
2080 - `enabled: boolean`
2081
2082 - `name: "edit"`
2083
2084 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2085
2086 Permission policy for tool execution.
2087
2088 - `BetaManagedAgentsAlwaysAllowPolicy object`
2089
2090 Tool calls are automatically approved without user confirmation.
2091
2092 - `BetaManagedAgentsAlwaysAskPolicy object`
2093
2094 Tool calls require user confirmation before execution.
2095
2096 - `BetaManagedAgentsAutoPolicy object`
2097
2098 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2099
2100 - `BetaManagedAgentsReadToolConfig object`
2101
2102 Configuration for the read tool.
2103
2104 - `type: "read"`
2105
2106 - `enabled: boolean`
2107
2108 - `name: "read"`
2109
2110 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2111
2112 Permission policy for tool execution.
2113
2114 - `BetaManagedAgentsAlwaysAllowPolicy object`
2115
2116 Tool calls are automatically approved without user confirmation.
2117
2118 - `BetaManagedAgentsAlwaysAskPolicy object`
2119
2120 Tool calls require user confirmation before execution.
2121
2122 - `BetaManagedAgentsAutoPolicy object`
2123
2124 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2125
2126 - `BetaManagedAgentsWriteToolConfig object`
2127
2128 Configuration for the write tool.
2129
2130 - `type: "write"`
2131
2132 - `enabled: boolean`
2133
2134 - `name: "write"`
2135
2136 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2137
2138 Permission policy for tool execution.
2139
2140 - `BetaManagedAgentsAlwaysAllowPolicy object`
2141
2142 Tool calls are automatically approved without user confirmation.
2143
2144 - `BetaManagedAgentsAlwaysAskPolicy object`
2145
2146 Tool calls require user confirmation before execution.
2147
2148 - `BetaManagedAgentsAutoPolicy object`
2149
2150 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2151
2152 - `BetaManagedAgentsGlobToolConfig object`
2153
2154 Configuration for the glob tool.
2155
2156 - `type: "glob"`
2157
2158 - `enabled: boolean`
2159
2160 - `name: "glob"`
2161
2162 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2163
2164 Permission policy for tool execution.
2165
2166 - `BetaManagedAgentsAlwaysAllowPolicy object`
2167
2168 Tool calls are automatically approved without user confirmation.
2169
2170 - `BetaManagedAgentsAlwaysAskPolicy object`
2171
2172 Tool calls require user confirmation before execution.
2173
2174 - `BetaManagedAgentsAutoPolicy object`
2175
2176 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2177
2178 - `BetaManagedAgentsGrepToolConfig object`
2179
2180 Configuration for the grep tool.
2181
2182 - `type: "grep"`
2183
2184 - `enabled: boolean`
2185
2186 - `name: "grep"`
2187
2188 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2189
2190 Permission policy for tool execution.
2191
2192 - `BetaManagedAgentsAlwaysAllowPolicy object`
2193
2194 Tool calls are automatically approved without user confirmation.
2195
2196 - `BetaManagedAgentsAlwaysAskPolicy object`
2197
2198 Tool calls require user confirmation before execution.
2199
2200 - `BetaManagedAgentsAutoPolicy object`
2201
2202 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2203
2204 - `BetaManagedAgentsWebFetchToolConfig object`
2205
2206 Configuration for the web_fetch tool.
2207
2208 - `type: "web_fetch"`
2209
2210 - `enabled: boolean`
2211
2212 - `name: "web_fetch"`
2213
2214 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2215
2216 Permission policy for tool execution.
2217
2218 - `BetaManagedAgentsAlwaysAllowPolicy object`
2219
2220 Tool calls are automatically approved without user confirmation.
2221
2222 - `BetaManagedAgentsAlwaysAskPolicy object`
2223
2224 Tool calls require user confirmation before execution.
2225
2226 - `BetaManagedAgentsAutoPolicy object`
2227
2228 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2229
2230 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
2231
2232 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
2233
2234 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2235
2236 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
2237
2238 - `BetaManagedAgentsWebFetchURLSourceAll object`
2239
2240 Every URL from this source may be fetched. This is the default.
2241
2242 - `type: "all"`
2243
2244 - `BetaManagedAgentsWebFetchURLSourceNone object`
2245
2246 This source contributes no URLs that may be fetched.
2247
2248 - `type: "none"`
2249
2250 - `BetaManagedAgentsWebFetchURLSourceOnly object`
2251
2252 Only the named tools' results contribute URLs that may be fetched.
2253
2254 - `type: "only"`
2255
2256 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2257
2258 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
2259
2260 - `type: "tool_reference"`
2261
2262 Must be "tool_reference".
2263
2264 - `name: string`
2265
2266 Name of the tool. Compared exactly, so upper and lower case letters are different.
2267
2268 minLength: 1, maxLength: 128
2269
2270 - `BetaManagedAgentsWebFetchURLSourceExcept object`
2271
2272 Every tool's results contribute URLs that may be fetched, except the named tools' results.
2273
2274 - `type: "except"`
2275
2276 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2277
2278 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
2279
2280 - `type: "tool_reference"`
2281
2282 Must be "tool_reference".
2283
2284 - `name: string`
2285
2286 Name of the tool. Compared exactly, so upper and lower case letters are different.
2287
2288 minLength: 1, maxLength: 128
2289
2290 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2291
2292 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
2293
2294 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
2295
2296 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
2297
2298 - `BetaManagedAgentsWebFetchURLSourceAll object`
2299
2300 Every URL from this source may be fetched. This is the default.
2301
2302 - `BetaManagedAgentsWebFetchURLSourceNone object`
2303
2304 This source contributes no URLs that may be fetched.
2305
2306 - `allowed_domains: optional array of string`
2307
2308 - `blocked_domains: optional array of string`
2309
2310 - `max_content_tokens: optional number or null`
2311
2312 format: int32
2313
2314 - `BetaManagedAgentsWebSearchToolConfig object`
2315
2316 Configuration for the web_search tool.
2317
2318 - `type: "web_search"`
2319
2320 - `enabled: boolean`
2321
2322 - `name: "web_search"`
2323
2324 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2325
2326 Permission policy for tool execution.
2327
2328 - `BetaManagedAgentsAlwaysAllowPolicy object`
2329
2330 Tool calls are automatically approved without user confirmation.
2331
2332 - `BetaManagedAgentsAlwaysAskPolicy object`
2333
2334 Tool calls require user confirmation before execution.
2335
2336 - `BetaManagedAgentsAutoPolicy object`
2337
2338 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2339
2340 - `allowed_domains: optional array of string`
2341
2342 - `blocked_domains: optional array of string`
2343
2344 - `user_location: optional BetaManagedAgentsUserLocation or null`
2345
2346 Approximate user location for search result localization.
2347
2348 - `type: "approximate"`
2349
2350 Location precision. Only "approximate" is supported.
2351
2352 - `city: optional string or null`
2353
2354 City name.
2355
2356 minLength: 1, maxLength: 255
2357
2358 - `country: optional string or null`
2359
2360 Two-letter ISO 3166-1 country code, uppercase.
2361
2362 - `region: optional string or null`
2363
2364 Region or state name.
2365
2366 minLength: 1, maxLength: 255
2367
2368 - `timezone: optional string or null`
2369
2370 IANA timezone identifier, e.g. "America/Los_Angeles".
2371
2372 minLength: 1, maxLength: 255
2373
2374 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
2375
2376 Resolved default configuration for agent tools.
20432377
20442378 - `enabled: boolean`
20452379
2046 - `name: "bash"`
2047
20482380 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
20492381
20502382 Permission policy for tool execution.
from line 2385
20532385
20542386 Tool calls are automatically approved without user confirmation.
20552387
2056 - `type: "always_allow"`
2057
20582388 - `BetaManagedAgentsAlwaysAskPolicy object`
20592389
20602390 Tool calls require user confirmation before execution.
20612391
2062 - `type: "always_ask"`
2063
20642392 - `BetaManagedAgentsAutoPolicy object`
20652393
20662394 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
20672395
2068 - `type: "auto"`
2069
2070 - `BetaManagedAgentsEditToolConfig object`
2071
2072 Configuration for the edit tool.
2073
2074 - `type: "edit"`
2396 - `BetaManagedAgentsMCPToolset object`
2397
2398 - `type: "mcp_toolset"`
2399
2400 - `configs: array of BetaManagedAgentsMCPToolConfig`
20752401
20762402 - `enabled: boolean`
20772403
2078 - `name: "edit"`
2404 - `name: string`
20792405
20802406 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
20812407
from line 2419
20932419
20942420 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
20952421
2096 - `BetaManagedAgentsReadToolConfig object`
2097
2098 Configuration for the read tool.
2099
2100 - `type: "read"`
2422 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
2423
2424 Resolved default configuration for all tools from an MCP server.
21012425
21022426 - `enabled: boolean`
21032427
2104 - `name: "read"`
2105
21062428 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
21072429
21082430 Permission policy for tool execution.
from line 2441
21192441
21202442 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
21212443
2122 - `BetaManagedAgentsWriteToolConfig object`
2123
2124 Configuration for the write tool.
2125
2126 - `type: "write"`
2127
2128 - `enabled: boolean`
2129
2130 - `name: "write"`
2131
2132 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2133
2134 Permission policy for tool execution.
2135
2136 - `BetaManagedAgentsAlwaysAllowPolicy object`
2137
2138 Tool calls are automatically approved without user confirmation.
2139
2140 - `BetaManagedAgentsAlwaysAskPolicy object`
2141
2142 Tool calls require user confirmation before execution.
2143
2144 - `BetaManagedAgentsAutoPolicy object`
2145
2146 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2147
2148 - `BetaManagedAgentsGlobToolConfig object`
2149
2150 Configuration for the glob tool.
2151
2152 - `type: "glob"`
2153
2154 - `enabled: boolean`
2155
2156 - `name: "glob"`
2157
2158 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2159
2160 Permission policy for tool execution.
2161
2162 - `BetaManagedAgentsAlwaysAllowPolicy object`
2163
2164 Tool calls are automatically approved without user confirmation.
2165
2166 - `BetaManagedAgentsAlwaysAskPolicy object`
2167
2168 Tool calls require user confirmation before execution.
2169
2170 - `BetaManagedAgentsAutoPolicy object`
2171
2172 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2173
2174 - `BetaManagedAgentsGrepToolConfig object`
2175
2176 Configuration for the grep tool.
2177
2178 - `type: "grep"`
2179
2180 - `enabled: boolean`
2181
2182 - `name: "grep"`
2183
2184 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2185
2186 Permission policy for tool execution.
2187
2188 - `BetaManagedAgentsAlwaysAllowPolicy object`
2189
2190 Tool calls are automatically approved without user confirmation.
2191
2192 - `BetaManagedAgentsAlwaysAskPolicy object`
2193
2194 Tool calls require user confirmation before execution.
2195
2196 - `BetaManagedAgentsAutoPolicy object`
2197
2198 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2199
2200 - `BetaManagedAgentsWebFetchToolConfig object`
2201
2202 Configuration for the web_fetch tool.
2203
2204 - `type: "web_fetch"`
2205
2206 - `enabled: boolean`
2207
2208 - `name: "web_fetch"`
2209
2210 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2211
2212 Permission policy for tool execution.
2213
2214 - `BetaManagedAgentsAlwaysAllowPolicy object`
2215
2216 Tool calls are automatically approved without user confirmation.
2217
2218 - `BetaManagedAgentsAlwaysAskPolicy object`
2219
2220 Tool calls require user confirmation before execution.
2221
2222 - `BetaManagedAgentsAutoPolicy object`
2223
2224 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2225
2226 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
2227
2228 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
2229
2230 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2231
2232 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
2233
2234 - `BetaManagedAgentsWebFetchURLSourceAll object`
2235
2236 Every URL from this source may be fetched. This is the default.
2237
2238 - `type: "all"`
2239
2240 - `BetaManagedAgentsWebFetchURLSourceNone object`
2241
2242 This source contributes no URLs that may be fetched.
2243
2244 - `type: "none"`
2245
2246 - `BetaManagedAgentsWebFetchURLSourceOnly object`
2247
2248 Only the named tools' results contribute URLs that may be fetched.
2249
2250 - `type: "only"`
2251
2252 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2253
2254 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
2255
2256 - `type: "tool_reference"`
2257
2258 Must be "tool_reference".
2259
2260 - `name: string`
2261
2262 Name of the tool. Compared exactly, so upper and lower case letters are different.
2263
2264 minLength: 1, maxLength: 128
2265
2266 - `BetaManagedAgentsWebFetchURLSourceExcept object`
2267
2268 Every tool's results contribute URLs that may be fetched, except the named tools' results.
2269
2270 - `type: "except"`
2271
2272 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
2273
2274 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
2275
2276 - `type: "tool_reference"`
2277
2278 Must be "tool_reference".
2279
2280 - `name: string`
2281
2282 Name of the tool. Compared exactly, so upper and lower case letters are different.
2283
2284 minLength: 1, maxLength: 128
2285
2286 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
2287
2288 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
2289
2290 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
2291
2292 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
2293
2294 - `BetaManagedAgentsWebFetchURLSourceAll object`
2295
2296 Every URL from this source may be fetched. This is the default.
2297
2298 - `BetaManagedAgentsWebFetchURLSourceNone object`
2299
2300 This source contributes no URLs that may be fetched.
2301
2302 - `allowed_domains: optional array of string`
2303
2304 - `blocked_domains: optional array of string`
2305
2306 - `max_content_tokens: optional number or null`
2307
2308 format: int32
2309
2310 - `BetaManagedAgentsWebSearchToolConfig object`
2311
2312 Configuration for the web_search tool.
2313
2314 - `type: "web_search"`
2315
2316 - `enabled: boolean`
2317
2318 - `name: "web_search"`
2319
2320 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2321
2322 Permission policy for tool execution.
2323
2324 - `BetaManagedAgentsAlwaysAllowPolicy object`
2325
2326 Tool calls are automatically approved without user confirmation.
2327
2328 - `BetaManagedAgentsAlwaysAskPolicy object`
2329
2330 Tool calls require user confirmation before execution.
2331
2332 - `BetaManagedAgentsAutoPolicy object`
2333
2334 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2335
2336 - `allowed_domains: optional array of string`
2337
2338 - `blocked_domains: optional array of string`
2339
2340 - `user_location: optional BetaManagedAgentsUserLocation or null`
2341
2342 Approximate user location for search result localization.
2343
2344 - `type: "approximate"`
2345
2346 Location precision. Only "approximate" is supported.
2347
2348 - `city: optional string or null`
2349
2350 City name.
2351
2352 minLength: 1, maxLength: 255
2353
2354 - `country: optional string or null`
2355
2356 Two-letter ISO 3166-1 country code, uppercase.
2357
2358 - `region: optional string or null`
2359
2360 Region or state name.
2361
2362 minLength: 1, maxLength: 255
2363
2364 - `timezone: optional string or null`
2365
2366 IANA timezone identifier, e.g. "America/Los_Angeles".
2367
2368 minLength: 1, maxLength: 255
2369
2370 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
2371
2372 Resolved default configuration for agent tools.
2373
2374 - `enabled: boolean`
2375
2376 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2377
2378 Permission policy for tool execution.
2379
2380 - `BetaManagedAgentsAlwaysAllowPolicy object`
2381
2382 Tool calls are automatically approved without user confirmation.
2383
2384 - `BetaManagedAgentsAlwaysAskPolicy object`
2385
2386 Tool calls require user confirmation before execution.
2387
2388 - `BetaManagedAgentsAutoPolicy object`
2389
2390 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2391
2392 - `BetaManagedAgentsMCPToolset object`
2393
2394 - `type: "mcp_toolset"`
2395
2396 - `configs: array of BetaManagedAgentsMCPToolConfig`
2397
2398 - `enabled: boolean`
2444 - `mcp_server_name: string`
2445
2446 - `BetaManagedAgentsCustomTool object`
2447
2448 A custom tool as returned in API responses.
2449
2450 - `type: "custom"`
2451
2452 - `description: string`
2453
2454 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
2455
2456 JSON Schema for custom tool input parameters.
2457
2458 - `type: "object"`
2459
2460 - `properties: optional map[unknown] or null`
2461
2462 - `required: optional array of string or null`
23992463
24002464 - `name: string`
24012465
2402 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2403
2404 Permission policy for tool execution.
2405
2406 - `BetaManagedAgentsAlwaysAllowPolicy object`
2407
2408 Tool calls are automatically approved without user confirmation.
2409
2410 - `BetaManagedAgentsAlwaysAskPolicy object`
2411
2412 Tool calls require user confirmation before execution.
2413
2414 - `BetaManagedAgentsAutoPolicy object`
2415
2416 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2417
2418 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
2419
2420 Resolved default configuration for all tools from an MCP server.
2421
2422 - `enabled: boolean`
2423
2424 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
2425
2426 Permission policy for tool execution.
2427
2428 - `BetaManagedAgentsAlwaysAllowPolicy object`
2429
2430 Tool calls are automatically approved without user confirmation.
2431
2432 - `BetaManagedAgentsAlwaysAskPolicy object`
2433
2434 Tool calls require user confirmation before execution.
2435
2436 - `BetaManagedAgentsAutoPolicy object`
2437
2438 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
2439
2440 - `mcp_server_name: string`
2441
2442 - `BetaManagedAgentsCustomTool object`
2443
2444 A custom tool as returned in API responses.
2445
2446 - `type: "custom"`
2447
2448 - `description: string`
2449
2450 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
2451
2452 JSON Schema for custom tool input parameters.
2453
2454 - `type: "object"`
2455
2456 - `properties: optional map[unknown] or null`
2457
2458 - `required: optional array of string or null`
2466 - `version: number`
2467
2468 format: int32
2469
2470 - `BetaManagedAgentsAdvisor object`
2471
2472 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
2473
2474 - `type: "advisor"`
2475
2476 - `model: string`
2477
2478 The advisor model id.
2479
2480 - `BetaManagedAgentsSessionMultiagent20261001 object`
2481
2482 Resolved multiagent configuration with three members, as copied to the `session` at creation.
2483
2484 - `type: "multiagent_20261001"`
2485
2486 - `advisor: BetaManagedAgentsMultiagentAdvisor`
2487
2488 Whether the session's primary thread can consult an advisor model.
2489
2490 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
2491
2492 The session's primary thread can consult `model` mid-turn.
2493
2494 - `type: "enabled"`
2495
2496 - `model: string`
2497
2498 The advisor model id.
2499
2500 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
2501
2502 The agent has no advisor.
2503
2504 - `type: "disabled"`
2505
2506 - `subagents: BetaManagedAgentsSessionMultiagentSubagents`
2507
2508 Whether the agent can spawn session threads.
2509
2510 - `BetaManagedAgentsSessionMultiagentSubagentsEnabled object`
2511
2512 The agent can spawn session threads.
2513
2514 - `type: "enabled"`
2515
2516 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
2517
2518 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
2519
2520 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
2521
2522 The agent can define inline agents.
2523
2524 - `type: "enabled"`
2525
2526 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
2527
2528 The agent cannot define inline agents.
2529
2530 - `type: "disabled"`
2531
2532 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
2533
2534 Full `agent` definitions of the predefined agents, which are saved agents that this agent can spawn as session threads.
2535
2536 - `type: "agent"`
2537
2538 - `id: string`
2539
2540 - `description: string or null`
2541
2542 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
2543
2544 - `model: BetaManagedAgentsModelConfig`
2545
2546 Model identifier and configuration.
24592547
24602548 - `name: string`
24612549
2462 - `version: number`
2463
2464 format: int32
2465
2466 - `BetaManagedAgentsAdvisor object`
2467
2468 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
2469
2470 - `type: "advisor"`
2471
2472 - `model: string`
2473
2474 The advisor model id.
2550 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2551
2552 - `system: string or null`
2553
2554 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2555
2556 - `version: number`
2557
2558 format: int32
2559
2560 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
2561
2562 The agent cannot spawn session threads.
2563
2564 - `type: "disabled"`
2565
2566 - `workflows: BetaManagedAgentsSessionMultiagentWorkflows`
2567
2568 Whether the agent can start workflow runs.
2569
2570 - `BetaManagedAgentsSessionMultiagentWorkflowsEnabled object`
2571
2572 The agent can start workflow runs.
2573
2574 - `type: "enabled"`
2575
2576 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
2577
2578 Whether a run's plan can define inline agents, which are not saved.
2579
2580 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
2581
2582 Full `agent` definitions of the predefined agents, which are saved agents that a run's plan can use.
2583
2584 - `type: "agent"`
2585
2586 - `id: string`
2587
2588 - `description: string or null`
2589
2590 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
2591
2592 - `model: BetaManagedAgentsModelConfig`
2593
2594 Model identifier and configuration.
2595
2596 - `name: string`
2597
2598 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
2599
2600 - `system: string or null`
2601
2602 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
2603
2604 - `version: number`
2605
2606 format: int32
2607
2608 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
2609
2610 The agent cannot start workflow runs.
2611
2612 - `type: "disabled"`
24752613
24762614 - `name: string`
24772615
from line 2826
26882826 - `budget: optional BetaManagedAgentsBudgetLimit or null`
26892827
26902828 The session's configured budget at the snapshot time, or null when the session has no budget.
2829
2830 - `BetaManagedAgentsWorkflowRunCreatedEvent object`
2831
2832 A workflow run was created. A workflow run is background work that the session's agent starts. Emitted once per run, before the run's other `workflow_run.*` events.
2833
2834 - `type: "workflow_run.created"`
2835
2836 - `id: string`
2837
2838 Unique identifier for this event.
2839
2840 - `description: string or null`
2841
2842 Description that the agent gave the run, passed on as written, or `null` if it gave none.
2843
2844 - `name: string`
2845
2846 Name that the agent gave the run, passed on as written, or a name that the server assigned.
2847
2848 - `phases: array of BetaManagedAgentsWorkflowRunPhase`
2849
2850 The phases that the run's plan declares, in the plan's order. Can be empty.
2851
2852 - `id: string`
2853
2854 Unique identifier for the phase.
2855
2856 - `description: string or null`
2857
2858 Description that the agent gave the phase, passed on as written, or `null` if it gave none.
2859
2860 - `name: string`
2861
2862 Name that the agent gave the phase, passed on as written.
2863
2864 - `processed_at: string`
2865
2866 Timestamp when this event was processed.
2867
2868 format: date-time
2869
2870 - `workflow_run_id: string`
2871
2872 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
2873
2874 - `BetaManagedAgentsWorkflowRunStatusEndedEvent object`
2875
2876 A workflow run ended. Emitted once per run, as the last of the run's `workflow_run.*` events.
2877
2878 - `type: "workflow_run.status_ended"`
2879
2880 - `id: string`
2881
2882 Unique identifier for this event.
2883
2884 - `processed_at: string`
2885
2886 Timestamp when this event was processed.
2887
2888 format: date-time
2889
2890 - `result: BetaManagedAgentsWorkflowRunResult`
2891
2892 How the run ended.
2893
2894 - `BetaManagedAgentsWorkflowRunResultCompleted object`
2895
2896 The run's plan, a program that the agent wrote, finished. This does not say whether the work succeeded.
2897
2898 - `type: "completed"`
2899
2900 - `BetaManagedAgentsWorkflowRunResultError object`
2901
2902 The run failed or reached its time limit.
2903
2904 - `type: "error"`
2905
2906 - `error: BetaManagedAgentsWorkflowRunError`
2907
2908 Why the run did not finish.
2909
2910 - `BetaManagedAgentsTimeoutWorkflowRunError object`
2911
2912 The run reached its time limit.
2913
2914 - `type: "timeout_error"`
2915
2916 - `message: string`
2917
2918 Short explanation written by the server. It never contains content from the run or its agents.
2919
2920 - `BetaManagedAgentsProgramWorkflowRunError object`
2921
2922 The plan, a program that the agent wrote, failed, or the server refused it.
2923
2924 - `type: "program_error"`
2925
2926 - `message: string`
2927
2928 Short explanation written by the server. It never contains content from the run or its agents.
2929
2930 - `BetaManagedAgentsUnknownWorkflowRunError object`
2931
2932 A failure that has no type of its own.
2933
2934 - `type: "unknown_error"`
2935
2936 - `message: string`
2937
2938 Short explanation written by the server. It never contains content from the run or its agents.
2939
2940 - `BetaManagedAgentsThreadLimitWorkflowRunError object`
2941
2942 The run exceeded the limit on the number of threads that a run can create.
2943
2944 - `type: "thread_limit_error"`
2945
2946 - `message: string`
2947
2948 Short explanation written by the server. It never contains content from the run or its agents.
2949
2950 - `BetaManagedAgentsMaxWorkflowRunsWorkflowRunError object`
2951
2952 No run was created, because the session was at its limit of open workflow runs, which are runs that have not ended. Only `workflow_run.error` carries this type.
2953
2954 - `type: "max_workflow_runs_error"`
2955
2956 - `message: string`
2957
2958 Short explanation written by the server. It never contains content from the run or its agents.
2959
2960 - `BetaManagedAgentsWorkflowRunResultStopped object`
2961
2962 The agent stopped the run.
2963
2964 - `type: "stopped"`
2965
2966 - `workflow_run_id: string`
2967
2968 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
2969
2970 - `BetaManagedAgentsWorkflowRunPhaseStartedEvent object`
2971
2972 A workflow run's plan entered a phase.
2973
2974 - `type: "workflow_run.phase_started"`
2975
2976 - `id: string`
2977
2978 Unique identifier for this event.
2979
2980 - `processed_at: string`
2981
2982 Timestamp when this event was processed.
2983
2984 format: date-time
2985
2986 - `workflow_run_id: string`
2987
2988 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
2989
2990 - `workflow_run_phase_id: string`
2991
2992 Identifier of the phase, as in `phases` on the run's `workflow_run.created` event.
2993
2994 - `BetaManagedAgentsWorkflowRunPhaseEndedEvent object`
2995
2996 A workflow run's plan left a phase, or the run's end closed it. Emitted once for every `workflow_run.phase_started` event, before the run's `workflow_run.status_ended` event. The event does not say whether the plan finished the phase's work, or why it left.
2997
2998 - `type: "workflow_run.phase_ended"`
2999
3000 - `id: string`
3001
3002 Unique identifier for this event.
3003
3004 - `phase_started_id: string`
3005
3006 Identifier of the `workflow_run.phase_started` event that opened the phase.
3007
3008 - `processed_at: string`
3009
3010 Timestamp when this event was processed.
3011
3012 format: date-time
3013
3014 - `workflow_run_id: string`
3015
3016 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3017
3018 - `workflow_run_phase_id: string`
3019
3020 Identifier of the phase, as in `phases` on the run's `workflow_run.created` event.
3021
3022 - `BetaManagedAgentsWorkflowRunStatusRunningEvent object`
3023
3024 A workflow run is running. Emitted when the run starts to execute, and each time it resumes after being idle. A run that starts idle emits `workflow_run.status_idle` first.
3025
3026 - `type: "workflow_run.status_running"`
3027
3028 - `id: string`
3029
3030 Unique identifier for this event.
3031
3032 - `processed_at: string`
3033
3034 Timestamp when this event was processed.
3035
3036 format: date-time
3037
3038 - `workflow_run_id: string`
3039
3040 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3041
3042 - `BetaManagedAgentsWorkflowRunStatusIdleEvent object`
3043
3044 A workflow run is idle. Emitted each time the run goes idle, whatever the cause. If the run ends while idle, no `workflow_run.status_running` comes between this event and its `workflow_run.status_ended`.
3045
3046 - `type: "workflow_run.status_idle"`
3047
3048 - `id: string`
3049
3050 Unique identifier for this event.
3051
3052 - `processed_at: string`
3053
3054 Timestamp when this event was processed.
3055
3056 format: date-time
3057
3058 - `workflow_run_id: string`
3059
3060 Identifier of the run. The same value is on all of the run's `workflow_run.*` events.
3061
3062 - `BetaManagedAgentsWorkflowRunErrorEvent object`
3063
3064 A workflow run met an error, or an error kept a run from being created. A run that ends with a `result.type` of `error` emits this event before its `workflow_run.status_ended`, with the same `error`.
3065
3066 - `type: "workflow_run.error"`
3067
3068 - `id: string`
3069
3070 Unique identifier for this event.
3071
3072 - `error: BetaManagedAgentsWorkflowRunError`
3073
3074 Why the run did not finish, or was not created.
3075
3076 - `processed_at: string`
3077
3078 Timestamp when this event was processed.
3079
3080 format: date-time
3081
3082 - `workflow_run_id: string or null`
3083
3084 Identifier of the run that met the error, or `null` when the error kept a run from being created.
26913085
26923086## Example
26933087
26943088
api/beta/sessions/list Changed · +571 / -437 lines
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 295
295295
296296 - `"claude-haiku-4-5"`
297297
298 Fastest model with near-frontier intelligence
299
300298 - `"claude-haiku-4-5-20251001"`
301299
302 Fastest model with near-frontier intelligence
303
304300 - `"claude-opus-4-5"`
305301
306302 Powerful intelligence for long-running agents and coding
from line 365
369365
370366 - `"fast"`
371367
372 - `multiagent: BetaManagedAgentsSessionMultiagentCoordinator or null`
368 - `multiagent: BetaManagedAgentsSessionMultiagent or null`
373369
374370 Resolved multiagent orchestration configuration. Null when the agent is single-threaded.
375371
376 - `type: "coordinator"`
377
378 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
379
380 Full `agent` definitions the coordinator may spawn as session threads.
381
382 - `BetaManagedAgentsSessionThreadAgent object`
383
384 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
385
386 - `type: "agent"`
387
388 - `id: string`
389
390 - `description: string or null`
391
392 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
393
394 - `type: "url"`
372 - `BetaManagedAgentsSessionMultiagentCoordinator object`
373
374 Resolved coordinator topology with full agent definitions for each roster member.
375
376 - `type: "coordinator"`
377
378 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
379
380 Full `agent` definitions the coordinator may spawn as session threads.
381
382 - `BetaManagedAgentsSessionThreadAgent object`
383
384 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
385
386 - `type: "agent"`
387
388 - `id: string`
389
390 - `description: string or null`
391
392 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
393
394 - `type: "url"`
395
396 - `name: string`
397
398 - `url: string`
399
400 - `model: BetaManagedAgentsModelConfig`
401
402 Model identifier and configuration.
395403
396404 - `name: string`
397405
398 - `url: string`
399
400 - `model: BetaManagedAgentsModelConfig`
401
402 Model identifier and configuration.
403
404 - `name: string`
405
406 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
407
408 - `BetaManagedAgentsAnthropicSkill object`
409
410 A resolved Anthropic-managed skill.
411
412 - `type: "anthropic"`
413
414 - `skill_id: string`
415
416 - `version: string`
417
418 - `BetaManagedAgentsCustomSkill object`
419
420 A resolved user-created custom skill.
421
422 - `type: "custom"`
423
424 - `skill_id: string`
425
426 - `version: string`
427
428 - `system: string or null`
429
430 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
431
432 - `BetaManagedAgentsAgentToolset20260401 object`
433
434 - `type: "agent_toolset_20260401"`
435
436 - `configs: array of BetaManagedAgentsAgentToolConfig`
437
438 - `BetaManagedAgentsBashToolConfig object`
439
440 Configuration for the bash tool.
441
442 - `type: "bash"`
406 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
407
408 - `BetaManagedAgentsAnthropicSkill object`
409
410 A resolved Anthropic-managed skill.
411
412 - `type: "anthropic"`
413
414 - `skill_id: string`
415
416 - `version: string`
417
418 - `BetaManagedAgentsCustomSkill object`
419
420 A resolved user-created custom skill.
421
422 - `type: "custom"`
423
424 - `skill_id: string`
425
426 - `version: string`
427
428 - `system: string or null`
429
430 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
431
432 - `BetaManagedAgentsAgentToolset20260401 object`
433
434 - `type: "agent_toolset_20260401"`
435
436 - `configs: array of BetaManagedAgentsAgentToolConfig`
437
438 - `BetaManagedAgentsBashToolConfig object`
439
440 Configuration for the bash tool.
441
442 - `type: "bash"`
443
444 - `enabled: boolean`
445
446 - `name: "bash"`
447
448 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
449
450 Permission policy for tool execution.
451
452 - `BetaManagedAgentsAlwaysAllowPolicy object`
453
454 Tool calls are automatically approved without user confirmation.
455
456 - `type: "always_allow"`
457
458 - `BetaManagedAgentsAlwaysAskPolicy object`
459
460 Tool calls require user confirmation before execution.
461
462 - `type: "always_ask"`
463
464 - `BetaManagedAgentsAutoPolicy object`
465
466 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
467
468 - `type: "auto"`
469
470 - `BetaManagedAgentsEditToolConfig object`
471
472 Configuration for the edit tool.
473
474 - `type: "edit"`
475
476 - `enabled: boolean`
477
478 - `name: "edit"`
479
480 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
481
482 Permission policy for tool execution.
483
484 - `BetaManagedAgentsAlwaysAllowPolicy object`
485
486 Tool calls are automatically approved without user confirmation.
487
488 - `BetaManagedAgentsAlwaysAskPolicy object`
489
490 Tool calls require user confirmation before execution.
491
492 - `BetaManagedAgentsAutoPolicy object`
493
494 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
495
496 - `BetaManagedAgentsReadToolConfig object`
497
498 Configuration for the read tool.
499
500 - `type: "read"`
501
502 - `enabled: boolean`
503
504 - `name: "read"`
505
506 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
507
508 Permission policy for tool execution.
509
510 - `BetaManagedAgentsAlwaysAllowPolicy object`
511
512 Tool calls are automatically approved without user confirmation.
513
514 - `BetaManagedAgentsAlwaysAskPolicy object`
515
516 Tool calls require user confirmation before execution.
517
518 - `BetaManagedAgentsAutoPolicy object`
519
520 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
521
522 - `BetaManagedAgentsWriteToolConfig object`
523
524 Configuration for the write tool.
525
526 - `type: "write"`
527
528 - `enabled: boolean`
529
530 - `name: "write"`
531
532 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
533
534 Permission policy for tool execution.
535
536 - `BetaManagedAgentsAlwaysAllowPolicy object`
537
538 Tool calls are automatically approved without user confirmation.
539
540 - `BetaManagedAgentsAlwaysAskPolicy object`
541
542 Tool calls require user confirmation before execution.
543
544 - `BetaManagedAgentsAutoPolicy object`
545
546 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
547
548 - `BetaManagedAgentsGlobToolConfig object`
549
550 Configuration for the glob tool.
551
552 - `type: "glob"`
553
554 - `enabled: boolean`
555
556 - `name: "glob"`
557
558 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
559
560 Permission policy for tool execution.
561
562 - `BetaManagedAgentsAlwaysAllowPolicy object`
563
564 Tool calls are automatically approved without user confirmation.
565
566 - `BetaManagedAgentsAlwaysAskPolicy object`
567
568 Tool calls require user confirmation before execution.
569
570 - `BetaManagedAgentsAutoPolicy object`
571
572 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
573
574 - `BetaManagedAgentsGrepToolConfig object`
575
576 Configuration for the grep tool.
577
578 - `type: "grep"`
579
580 - `enabled: boolean`
581
582 - `name: "grep"`
583
584 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
585
586 Permission policy for tool execution.
587
588 - `BetaManagedAgentsAlwaysAllowPolicy object`
589
590 Tool calls are automatically approved without user confirmation.
591
592 - `BetaManagedAgentsAlwaysAskPolicy object`
593
594 Tool calls require user confirmation before execution.
595
596 - `BetaManagedAgentsAutoPolicy object`
597
598 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
599
600 - `BetaManagedAgentsWebFetchToolConfig object`
601
602 Configuration for the web_fetch tool.
603
604 - `type: "web_fetch"`
605
606 - `enabled: boolean`
607
608 - `name: "web_fetch"`
609
610 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
611
612 Permission policy for tool execution.
613
614 - `BetaManagedAgentsAlwaysAllowPolicy object`
615
616 Tool calls are automatically approved without user confirmation.
617
618 - `BetaManagedAgentsAlwaysAskPolicy object`
619
620 Tool calls require user confirmation before execution.
621
622 - `BetaManagedAgentsAutoPolicy object`
623
624 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
625
626 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
627
628 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
629
630 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
631
632 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
633
634 - `BetaManagedAgentsWebFetchURLSourceAll object`
635
636 Every URL from this source may be fetched. This is the default.
637
638 - `type: "all"`
639
640 - `BetaManagedAgentsWebFetchURLSourceNone object`
641
642 This source contributes no URLs that may be fetched.
643
644 - `type: "none"`
645
646 - `BetaManagedAgentsWebFetchURLSourceOnly object`
647
648 Only the named tools' results contribute URLs that may be fetched.
649
650 - `type: "only"`
651
652 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
653
654 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
655
656 - `type: "tool_reference"`
657
658 Must be "tool_reference".
659
660 - `name: string`
661
662 Name of the tool. Compared exactly, so upper and lower case letters are different.
663
664 minLength: 1, maxLength: 128
665
666 - `BetaManagedAgentsWebFetchURLSourceExcept object`
667
668 Every tool's results contribute URLs that may be fetched, except the named tools' results.
669
670 - `type: "except"`
671
672 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
673
674 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
675
676 - `type: "tool_reference"`
677
678 Must be "tool_reference".
679
680 - `name: string`
681
682 Name of the tool. Compared exactly, so upper and lower case letters are different.
683
684 minLength: 1, maxLength: 128
685
686 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
687
688 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
689
690 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
691
692 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
693
694 - `BetaManagedAgentsWebFetchURLSourceAll object`
695
696 Every URL from this source may be fetched. This is the default.
697
698 - `BetaManagedAgentsWebFetchURLSourceNone object`
699
700 This source contributes no URLs that may be fetched.
701
702 - `allowed_domains: optional array of string`
703
704 - `blocked_domains: optional array of string`
705
706 - `max_content_tokens: optional number or null`
707
708 format: int32
709
710 - `BetaManagedAgentsWebSearchToolConfig object`
711
712 Configuration for the web_search tool.
713
714 - `type: "web_search"`
715
716 - `enabled: boolean`
717
718 - `name: "web_search"`
719
720 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
721
722 Permission policy for tool execution.
723
724 - `BetaManagedAgentsAlwaysAllowPolicy object`
725
726 Tool calls are automatically approved without user confirmation.
727
728 - `BetaManagedAgentsAlwaysAskPolicy object`
729
730 Tool calls require user confirmation before execution.
731
732 - `BetaManagedAgentsAutoPolicy object`
733
734 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
735
736 - `allowed_domains: optional array of string`
737
738 - `blocked_domains: optional array of string`
739
740 - `user_location: optional BetaManagedAgentsUserLocation or null`
741
742 Approximate user location for search result localization.
743
744 - `type: "approximate"`
745
746 Location precision. Only "approximate" is supported.
747
748 - `city: optional string or null`
749
750 City name.
751
752 minLength: 1, maxLength: 255
753
754 - `country: optional string or null`
755
756 Two-letter ISO 3166-1 country code, uppercase.
757
758 - `region: optional string or null`
759
760 Region or state name.
761
762 minLength: 1, maxLength: 255
763
764 - `timezone: optional string or null`
765
766 IANA timezone identifier, e.g. "America/Los_Angeles".
767
768 minLength: 1, maxLength: 255
769
770 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
771
772 Resolved default configuration for agent tools.
443773
444774 - `enabled: boolean`
445775
446 - `name: "bash"`
447
448776 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
449777
450778 Permission policy for tool execution.
from line 781
453781
454782 Tool calls are automatically approved without user confirmation.
455783
456 - `type: "always_allow"`
457
458784 - `BetaManagedAgentsAlwaysAskPolicy object`
459785
460786 Tool calls require user confirmation before execution.
461787
462 - `type: "always_ask"`
463
464788 - `BetaManagedAgentsAutoPolicy object`
465789
466790 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
467791
468 - `type: "auto"`
469
470 - `BetaManagedAgentsEditToolConfig object`
471
472 Configuration for the edit tool.
473
474 - `type: "edit"`
792 - `BetaManagedAgentsMCPToolset object`
793
794 - `type: "mcp_toolset"`
795
796 - `configs: array of BetaManagedAgentsMCPToolConfig`
475797
476798 - `enabled: boolean`
477799
478 - `name: "edit"`
800 - `name: string`
479801
480802 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
481803
from line 815
493815
494816 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
495817
496 - `BetaManagedAgentsReadToolConfig object`
497
498 Configuration for the read tool.
499
500 - `type: "read"`
818 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
819
820 Resolved default configuration for all tools from an MCP server.
501821
502822 - `enabled: boolean`
503823
504 - `name: "read"`
505
506824 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
507825
508826 Permission policy for tool execution.
from line 837
519837
520838 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
521839
522 - `BetaManagedAgentsWriteToolConfig object`
523
524 Configuration for the write tool.
525
526 - `type: "write"`
527
528 - `enabled: boolean`
529
530 - `name: "write"`
531
532 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
533
534 Permission policy for tool execution.
535
536 - `BetaManagedAgentsAlwaysAllowPolicy object`
537
538 Tool calls are automatically approved without user confirmation.
539
540 - `BetaManagedAgentsAlwaysAskPolicy object`
541
542 Tool calls require user confirmation before execution.
543
544 - `BetaManagedAgentsAutoPolicy object`
545
546 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
547
548 - `BetaManagedAgentsGlobToolConfig object`
549
550 Configuration for the glob tool.
551
552 - `type: "glob"`
553
554 - `enabled: boolean`
555
556 - `name: "glob"`
557
558 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
559
560 Permission policy for tool execution.
561
562 - `BetaManagedAgentsAlwaysAllowPolicy object`
563
564 Tool calls are automatically approved without user confirmation.
565
566 - `BetaManagedAgentsAlwaysAskPolicy object`
567
568 Tool calls require user confirmation before execution.
569
570 - `BetaManagedAgentsAutoPolicy object`
571
572 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
573
574 - `BetaManagedAgentsGrepToolConfig object`
575
576 Configuration for the grep tool.
577
578 - `type: "grep"`
579
580 - `enabled: boolean`
581
582 - `name: "grep"`
583
584 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
585
586 Permission policy for tool execution.
587
588 - `BetaManagedAgentsAlwaysAllowPolicy object`
589
590 Tool calls are automatically approved without user confirmation.
591
592 - `BetaManagedAgentsAlwaysAskPolicy object`
593
594 Tool calls require user confirmation before execution.
595
596 - `BetaManagedAgentsAutoPolicy object`
597
598 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
599
600 - `BetaManagedAgentsWebFetchToolConfig object`
601
602 Configuration for the web_fetch tool.
603
604 - `type: "web_fetch"`
605
606 - `enabled: boolean`
607
608 - `name: "web_fetch"`
609
610 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
611
612 Permission policy for tool execution.
613
614 - `BetaManagedAgentsAlwaysAllowPolicy object`
615
616 Tool calls are automatically approved without user confirmation.
617
618 - `BetaManagedAgentsAlwaysAskPolicy object`
619
620 Tool calls require user confirmation before execution.
621
622 - `BetaManagedAgentsAutoPolicy object`
623
624 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
625
626 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
627
628 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
629
630 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
631
632 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
633
634 - `BetaManagedAgentsWebFetchURLSourceAll object`
635
636 Every URL from this source may be fetched. This is the default.
637
638 - `type: "all"`
639
640 - `BetaManagedAgentsWebFetchURLSourceNone object`
641
642 This source contributes no URLs that may be fetched.
643
644 - `type: "none"`
645
646 - `BetaManagedAgentsWebFetchURLSourceOnly object`
647
648 Only the named tools' results contribute URLs that may be fetched.
649
650 - `type: "only"`
651
652 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
653
654 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
655
656 - `type: "tool_reference"`
657
658 Must be "tool_reference".
659
660 - `name: string`
661
662 Name of the tool. Compared exactly, so upper and lower case letters are different.
663
664 minLength: 1, maxLength: 128
665
666 - `BetaManagedAgentsWebFetchURLSourceExcept object`
667
668 Every tool's results contribute URLs that may be fetched, except the named tools' results.
669
670 - `type: "except"`
671
672 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
673
674 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
675
676 - `type: "tool_reference"`
677
678 Must be "tool_reference".
679
680 - `name: string`
681
682 Name of the tool. Compared exactly, so upper and lower case letters are different.
683
684 minLength: 1, maxLength: 128
685
686 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
687
688 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
689
690 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
691
692 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
693
694 - `BetaManagedAgentsWebFetchURLSourceAll object`
695
696 Every URL from this source may be fetched. This is the default.
697
698 - `BetaManagedAgentsWebFetchURLSourceNone object`
699
700 This source contributes no URLs that may be fetched.
701
702 - `allowed_domains: optional array of string`
703
704 - `blocked_domains: optional array of string`
705
706 - `max_content_tokens: optional number or null`
707
708 format: int32
709
710 - `BetaManagedAgentsWebSearchToolConfig object`
711
712 Configuration for the web_search tool.
713
714 - `type: "web_search"`
715
716 - `enabled: boolean`
717
718 - `name: "web_search"`
719
720 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
721
722 Permission policy for tool execution.
723
724 - `BetaManagedAgentsAlwaysAllowPolicy object`
725
726 Tool calls are automatically approved without user confirmation.
727
728 - `BetaManagedAgentsAlwaysAskPolicy object`
729
730 Tool calls require user confirmation before execution.
731
732 - `BetaManagedAgentsAutoPolicy object`
733
734 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
735
736 - `allowed_domains: optional array of string`
737
738 - `blocked_domains: optional array of string`
739
740 - `user_location: optional BetaManagedAgentsUserLocation or null`
741
742 Approximate user location for search result localization.
743
744 - `type: "approximate"`
745
746 Location precision. Only "approximate" is supported.
747
748 - `city: optional string or null`
749
750 City name.
751
752 minLength: 1, maxLength: 255
753
754 - `country: optional string or null`
755
756 Two-letter ISO 3166-1 country code, uppercase.
757
758 - `region: optional string or null`
759
760 Region or state name.
761
762 minLength: 1, maxLength: 255
763
764 - `timezone: optional string or null`
765
766 IANA timezone identifier, e.g. "America/Los_Angeles".
767
768 minLength: 1, maxLength: 255
769
770 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
771
772 Resolved default configuration for agent tools.
773
774 - `enabled: boolean`
775
776 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
777
778 Permission policy for tool execution.
779
780 - `BetaManagedAgentsAlwaysAllowPolicy object`
781
782 Tool calls are automatically approved without user confirmation.
783
784 - `BetaManagedAgentsAlwaysAskPolicy object`
785
786 Tool calls require user confirmation before execution.
787
788 - `BetaManagedAgentsAutoPolicy object`
789
790 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
791
792 - `BetaManagedAgentsMCPToolset object`
793
794 - `type: "mcp_toolset"`
795
796 - `configs: array of BetaManagedAgentsMCPToolConfig`
797
798 - `enabled: boolean`
840 - `mcp_server_name: string`
841
842 - `BetaManagedAgentsCustomTool object`
843
844 A custom tool as returned in API responses.
845
846 - `type: "custom"`
847
848 - `description: string`
849
850 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
851
852 JSON Schema for custom tool input parameters.
853
854 - `type: "object"`
855
856 - `properties: optional map[unknown] or null`
857
858 - `required: optional array of string or null`
799859
800860 - `name: string`
801861
802 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
803
804 Permission policy for tool execution.
805
806 - `BetaManagedAgentsAlwaysAllowPolicy object`
807
808 Tool calls are automatically approved without user confirmation.
809
810 - `BetaManagedAgentsAlwaysAskPolicy object`
811
812 Tool calls require user confirmation before execution.
813
814 - `BetaManagedAgentsAutoPolicy object`
815
816 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
817
818 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
819
820 Resolved default configuration for all tools from an MCP server.
821
822 - `enabled: boolean`
823
824 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
825
826 Permission policy for tool execution.
827
828 - `BetaManagedAgentsAlwaysAllowPolicy object`
829
830 Tool calls are automatically approved without user confirmation.
831
832 - `BetaManagedAgentsAlwaysAskPolicy object`
833
834 Tool calls require user confirmation before execution.
835
836 - `BetaManagedAgentsAutoPolicy object`
837
838 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
839
840 - `mcp_server_name: string`
841
842 - `BetaManagedAgentsCustomTool object`
843
844 A custom tool as returned in API responses.
845
846 - `type: "custom"`
847
848 - `description: string`
849
850 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
851
852 JSON Schema for custom tool input parameters.
853
854 - `type: "object"`
855
856 - `properties: optional map[unknown] or null`
857
858 - `required: optional array of string or null`
862 - `version: number`
863
864 format: int32
865
866 - `BetaManagedAgentsAdvisor object`
867
868 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
869
870 - `type: "advisor"`
871
872 - `model: string`
873
874 The advisor model id.
875
876 - `BetaManagedAgentsSessionMultiagent20261001 object`
877
878 Resolved multiagent configuration with three members, as copied to the `session` at creation.
879
880 - `type: "multiagent_20261001"`
881
882 - `advisor: BetaManagedAgentsMultiagentAdvisor`
883
884 Whether the session's primary thread can consult an advisor model.
885
886 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
887
888 The session's primary thread can consult `model` mid-turn.
889
890 - `type: "enabled"`
891
892 - `model: string`
893
894 The advisor model id.
895
896 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
897
898 The agent has no advisor.
899
900 - `type: "disabled"`
901
902 - `subagents: BetaManagedAgentsSessionMultiagentSubagents`
903
904 Whether the agent can spawn session threads.
905
906 - `BetaManagedAgentsSessionMultiagentSubagentsEnabled object`
907
908 The agent can spawn session threads.
909
910 - `type: "enabled"`
911
912 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
913
914 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
915
916 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
917
918 The agent can define inline agents.
919
920 - `type: "enabled"`
921
922 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
923
924 The agent cannot define inline agents.
925
926 - `type: "disabled"`
927
928 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
929
930 Full `agent` definitions of the predefined agents, which are saved agents that this agent can spawn as session threads.
931
932 - `type: "agent"`
933
934 - `id: string`
935
936 - `description: string or null`
937
938 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
939
940 - `model: BetaManagedAgentsModelConfig`
941
942 Model identifier and configuration.
859943
860944 - `name: string`
861945
862 - `version: number`
863
864 format: int32
865
866 - `BetaManagedAgentsAdvisor object`
867
868 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
869
870 - `type: "advisor"`
871
872 - `model: string`
873
874 The advisor model id.
946 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
947
948 - `system: string or null`
949
950 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
951
952 - `version: number`
953
954 format: int32
955
956 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
957
958 The agent cannot spawn session threads.
959
960 - `type: "disabled"`
961
962 - `workflows: BetaManagedAgentsSessionMultiagentWorkflows`
963
964 Whether the agent can start workflow runs.
965
966 - `BetaManagedAgentsSessionMultiagentWorkflowsEnabled object`
967
968 The agent can start workflow runs.
969
970 - `type: "enabled"`
971
972 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
973
974 Whether a run's plan can define inline agents, which are not saved.
975
976 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
977
978 Full `agent` definitions of the predefined agents, which are saved agents that a run's plan can use.
979
980 - `type: "agent"`
981
982 - `id: string`
983
984 - `description: string or null`
985
986 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
987
988 - `model: BetaManagedAgentsModelConfig`
989
990 Model identifier and configuration.
991
992 - `name: string`
993
994 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
995
996 - `system: string or null`
997
998 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
999
1000 - `version: number`
1001
1002 format: int32
1003
1004 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
1005
1006 The agent cannot start workflow runs.
1007
1008 - `type: "disabled"`
8751009
8761010 - `name: string`
8771011
8781012
api/beta/sessions/retrieve Changed · +571 / -437 lines
The two sides of this change are more than 400 edits apart, too far apart to line up, so this is the differ's own diff of it and the words inside a line are not marked.
from line 213
213213
214214 - `"claude-haiku-4-5"`
215215
216 Fastest model with near-frontier intelligence
217
218216 - `"claude-haiku-4-5-20251001"`
219217
220 Fastest model with near-frontier intelligence
221
222218 - `"claude-opus-4-5"`
223219
224220 Powerful intelligence for long-running agents and coding
from line 283
287283
288284 - `"fast"`
289285
290 - `multiagent: BetaManagedAgentsSessionMultiagentCoordinator or null`
286 - `multiagent: BetaManagedAgentsSessionMultiagent or null`
291287
292288 Resolved multiagent orchestration configuration. Null when the agent is single-threaded.
293289
294 - `type: "coordinator"`
295
296 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
297
298 Full `agent` definitions the coordinator may spawn as session threads.
299
300 - `BetaManagedAgentsSessionThreadAgent object`
301
302 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
303
304 - `type: "agent"`
305
306 - `id: string`
307
308 - `description: string or null`
309
310 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
311
312 - `type: "url"`
290 - `BetaManagedAgentsSessionMultiagentCoordinator object`
291
292 Resolved coordinator topology with full agent definitions for each roster member.
293
294 - `type: "coordinator"`
295
296 - `agents: array of BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
297
298 Full `agent` definitions the coordinator may spawn as session threads.
299
300 - `BetaManagedAgentsSessionThreadAgent object`
301
302 Resolved `agent` definition for a single `session_thread`. Snapshot of the agent at thread creation time. The multiagent roster is not repeated here; read it from `Session.agent`.
303
304 - `type: "agent"`
305
306 - `id: string`
307
308 - `description: string or null`
309
310 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
311
312 - `type: "url"`
313
314 - `name: string`
315
316 - `url: string`
317
318 - `model: BetaManagedAgentsModelConfig`
319
320 Model identifier and configuration.
313321
314322 - `name: string`
315323
316 - `url: string`
317
318 - `model: BetaManagedAgentsModelConfig`
319
320 Model identifier and configuration.
321
322 - `name: string`
323
324 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
325
326 - `BetaManagedAgentsAnthropicSkill object`
327
328 A resolved Anthropic-managed skill.
329
330 - `type: "anthropic"`
331
332 - `skill_id: string`
333
334 - `version: string`
335
336 - `BetaManagedAgentsCustomSkill object`
337
338 A resolved user-created custom skill.
339
340 - `type: "custom"`
341
342 - `skill_id: string`
343
344 - `version: string`
345
346 - `system: string or null`
347
348 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
349
350 - `BetaManagedAgentsAgentToolset20260401 object`
351
352 - `type: "agent_toolset_20260401"`
353
354 - `configs: array of BetaManagedAgentsAgentToolConfig`
355
356 - `BetaManagedAgentsBashToolConfig object`
357
358 Configuration for the bash tool.
359
360 - `type: "bash"`
324 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
325
326 - `BetaManagedAgentsAnthropicSkill object`
327
328 A resolved Anthropic-managed skill.
329
330 - `type: "anthropic"`
331
332 - `skill_id: string`
333
334 - `version: string`
335
336 - `BetaManagedAgentsCustomSkill object`
337
338 A resolved user-created custom skill.
339
340 - `type: "custom"`
341
342 - `skill_id: string`
343
344 - `version: string`
345
346 - `system: string or null`
347
348 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
349
350 - `BetaManagedAgentsAgentToolset20260401 object`
351
352 - `type: "agent_toolset_20260401"`
353
354 - `configs: array of BetaManagedAgentsAgentToolConfig`
355
356 - `BetaManagedAgentsBashToolConfig object`
357
358 Configuration for the bash tool.
359
360 - `type: "bash"`
361
362 - `enabled: boolean`
363
364 - `name: "bash"`
365
366 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
367
368 Permission policy for tool execution.
369
370 - `BetaManagedAgentsAlwaysAllowPolicy object`
371
372 Tool calls are automatically approved without user confirmation.
373
374 - `type: "always_allow"`
375
376 - `BetaManagedAgentsAlwaysAskPolicy object`
377
378 Tool calls require user confirmation before execution.
379
380 - `type: "always_ask"`
381
382 - `BetaManagedAgentsAutoPolicy object`
383
384 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
385
386 - `type: "auto"`
387
388 - `BetaManagedAgentsEditToolConfig object`
389
390 Configuration for the edit tool.
391
392 - `type: "edit"`
393
394 - `enabled: boolean`
395
396 - `name: "edit"`
397
398 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
399
400 Permission policy for tool execution.
401
402 - `BetaManagedAgentsAlwaysAllowPolicy object`
403
404 Tool calls are automatically approved without user confirmation.
405
406 - `BetaManagedAgentsAlwaysAskPolicy object`
407
408 Tool calls require user confirmation before execution.
409
410 - `BetaManagedAgentsAutoPolicy object`
411
412 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
413
414 - `BetaManagedAgentsReadToolConfig object`
415
416 Configuration for the read tool.
417
418 - `type: "read"`
419
420 - `enabled: boolean`
421
422 - `name: "read"`
423
424 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
425
426 Permission policy for tool execution.
427
428 - `BetaManagedAgentsAlwaysAllowPolicy object`
429
430 Tool calls are automatically approved without user confirmation.
431
432 - `BetaManagedAgentsAlwaysAskPolicy object`
433
434 Tool calls require user confirmation before execution.
435
436 - `BetaManagedAgentsAutoPolicy object`
437
438 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
439
440 - `BetaManagedAgentsWriteToolConfig object`
441
442 Configuration for the write tool.
443
444 - `type: "write"`
445
446 - `enabled: boolean`
447
448 - `name: "write"`
449
450 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
451
452 Permission policy for tool execution.
453
454 - `BetaManagedAgentsAlwaysAllowPolicy object`
455
456 Tool calls are automatically approved without user confirmation.
457
458 - `BetaManagedAgentsAlwaysAskPolicy object`
459
460 Tool calls require user confirmation before execution.
461
462 - `BetaManagedAgentsAutoPolicy object`
463
464 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
465
466 - `BetaManagedAgentsGlobToolConfig object`
467
468 Configuration for the glob tool.
469
470 - `type: "glob"`
471
472 - `enabled: boolean`
473
474 - `name: "glob"`
475
476 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
477
478 Permission policy for tool execution.
479
480 - `BetaManagedAgentsAlwaysAllowPolicy object`
481
482 Tool calls are automatically approved without user confirmation.
483
484 - `BetaManagedAgentsAlwaysAskPolicy object`
485
486 Tool calls require user confirmation before execution.
487
488 - `BetaManagedAgentsAutoPolicy object`
489
490 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
491
492 - `BetaManagedAgentsGrepToolConfig object`
493
494 Configuration for the grep tool.
495
496 - `type: "grep"`
497
498 - `enabled: boolean`
499
500 - `name: "grep"`
501
502 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
503
504 Permission policy for tool execution.
505
506 - `BetaManagedAgentsAlwaysAllowPolicy object`
507
508 Tool calls are automatically approved without user confirmation.
509
510 - `BetaManagedAgentsAlwaysAskPolicy object`
511
512 Tool calls require user confirmation before execution.
513
514 - `BetaManagedAgentsAutoPolicy object`
515
516 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
517
518 - `BetaManagedAgentsWebFetchToolConfig object`
519
520 Configuration for the web_fetch tool.
521
522 - `type: "web_fetch"`
523
524 - `enabled: boolean`
525
526 - `name: "web_fetch"`
527
528 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
529
530 Permission policy for tool execution.
531
532 - `BetaManagedAgentsAlwaysAllowPolicy object`
533
534 Tool calls are automatically approved without user confirmation.
535
536 - `BetaManagedAgentsAlwaysAskPolicy object`
537
538 Tool calls require user confirmation before execution.
539
540 - `BetaManagedAgentsAutoPolicy object`
541
542 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
543
544 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
545
546 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
547
548 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
549
550 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
551
552 - `BetaManagedAgentsWebFetchURLSourceAll object`
553
554 Every URL from this source may be fetched. This is the default.
555
556 - `type: "all"`
557
558 - `BetaManagedAgentsWebFetchURLSourceNone object`
559
560 This source contributes no URLs that may be fetched.
561
562 - `type: "none"`
563
564 - `BetaManagedAgentsWebFetchURLSourceOnly object`
565
566 Only the named tools' results contribute URLs that may be fetched.
567
568 - `type: "only"`
569
570 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
571
572 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
573
574 - `type: "tool_reference"`
575
576 Must be "tool_reference".
577
578 - `name: string`
579
580 Name of the tool. Compared exactly, so upper and lower case letters are different.
581
582 minLength: 1, maxLength: 128
583
584 - `BetaManagedAgentsWebFetchURLSourceExcept object`
585
586 Every tool's results contribute URLs that may be fetched, except the named tools' results.
587
588 - `type: "except"`
589
590 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
591
592 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
593
594 - `type: "tool_reference"`
595
596 Must be "tool_reference".
597
598 - `name: string`
599
600 Name of the tool. Compared exactly, so upper and lower case letters are different.
601
602 minLength: 1, maxLength: 128
603
604 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
605
606 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
607
608 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
609
610 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
611
612 - `BetaManagedAgentsWebFetchURLSourceAll object`
613
614 Every URL from this source may be fetched. This is the default.
615
616 - `BetaManagedAgentsWebFetchURLSourceNone object`
617
618 This source contributes no URLs that may be fetched.
619
620 - `allowed_domains: optional array of string`
621
622 - `blocked_domains: optional array of string`
623
624 - `max_content_tokens: optional number or null`
625
626 format: int32
627
628 - `BetaManagedAgentsWebSearchToolConfig object`
629
630 Configuration for the web_search tool.
631
632 - `type: "web_search"`
633
634 - `enabled: boolean`
635
636 - `name: "web_search"`
637
638 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
639
640 Permission policy for tool execution.
641
642 - `BetaManagedAgentsAlwaysAllowPolicy object`
643
644 Tool calls are automatically approved without user confirmation.
645
646 - `BetaManagedAgentsAlwaysAskPolicy object`
647
648 Tool calls require user confirmation before execution.
649
650 - `BetaManagedAgentsAutoPolicy object`
651
652 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
653
654 - `allowed_domains: optional array of string`
655
656 - `blocked_domains: optional array of string`
657
658 - `user_location: optional BetaManagedAgentsUserLocation or null`
659
660 Approximate user location for search result localization.
661
662 - `type: "approximate"`
663
664 Location precision. Only "approximate" is supported.
665
666 - `city: optional string or null`
667
668 City name.
669
670 minLength: 1, maxLength: 255
671
672 - `country: optional string or null`
673
674 Two-letter ISO 3166-1 country code, uppercase.
675
676 - `region: optional string or null`
677
678 Region or state name.
679
680 minLength: 1, maxLength: 255
681
682 - `timezone: optional string or null`
683
684 IANA timezone identifier, e.g. "America/Los_Angeles".
685
686 minLength: 1, maxLength: 255
687
688 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
689
690 Resolved default configuration for agent tools.
361691
362692 - `enabled: boolean`
363693
364 - `name: "bash"`
365
366694 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
367695
368696 Permission policy for tool execution.
from line 699
371699
372700 Tool calls are automatically approved without user confirmation.
373701
374 - `type: "always_allow"`
375
376702 - `BetaManagedAgentsAlwaysAskPolicy object`
377703
378704 Tool calls require user confirmation before execution.
379705
380 - `type: "always_ask"`
381
382706 - `BetaManagedAgentsAutoPolicy object`
383707
384708 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
385709
386 - `type: "auto"`
387
388 - `BetaManagedAgentsEditToolConfig object`
389
390 Configuration for the edit tool.
391
392 - `type: "edit"`
710 - `BetaManagedAgentsMCPToolset object`
711
712 - `type: "mcp_toolset"`
713
714 - `configs: array of BetaManagedAgentsMCPToolConfig`
393715
394716 - `enabled: boolean`
395717
396 - `name: "edit"`
718 - `name: string`
397719
398720 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
399721
from line 733
411733
412734 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
413735
414 - `BetaManagedAgentsReadToolConfig object`
415
416 Configuration for the read tool.
417
418 - `type: "read"`
736 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
737
738 Resolved default configuration for all tools from an MCP server.
419739
420740 - `enabled: boolean`
421741
422 - `name: "read"`
423
424742 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
425743
426744 Permission policy for tool execution.
from line 755
437755
438756 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
439757
440 - `BetaManagedAgentsWriteToolConfig object`
441
442 Configuration for the write tool.
443
444 - `type: "write"`
445
446 - `enabled: boolean`
447
448 - `name: "write"`
449
450 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
451
452 Permission policy for tool execution.
453
454 - `BetaManagedAgentsAlwaysAllowPolicy object`
455
456 Tool calls are automatically approved without user confirmation.
457
458 - `BetaManagedAgentsAlwaysAskPolicy object`
459
460 Tool calls require user confirmation before execution.
461
462 - `BetaManagedAgentsAutoPolicy object`
463
464 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
465
466 - `BetaManagedAgentsGlobToolConfig object`
467
468 Configuration for the glob tool.
469
470 - `type: "glob"`
471
472 - `enabled: boolean`
473
474 - `name: "glob"`
475
476 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
477
478 Permission policy for tool execution.
479
480 - `BetaManagedAgentsAlwaysAllowPolicy object`
481
482 Tool calls are automatically approved without user confirmation.
483
484 - `BetaManagedAgentsAlwaysAskPolicy object`
485
486 Tool calls require user confirmation before execution.
487
488 - `BetaManagedAgentsAutoPolicy object`
489
490 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
491
492 - `BetaManagedAgentsGrepToolConfig object`
493
494 Configuration for the grep tool.
495
496 - `type: "grep"`
497
498 - `enabled: boolean`
499
500 - `name: "grep"`
501
502 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
503
504 Permission policy for tool execution.
505
506 - `BetaManagedAgentsAlwaysAllowPolicy object`
507
508 Tool calls are automatically approved without user confirmation.
509
510 - `BetaManagedAgentsAlwaysAskPolicy object`
511
512 Tool calls require user confirmation before execution.
513
514 - `BetaManagedAgentsAutoPolicy object`
515
516 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
517
518 - `BetaManagedAgentsWebFetchToolConfig object`
519
520 Configuration for the web_fetch tool.
521
522 - `type: "web_fetch"`
523
524 - `enabled: boolean`
525
526 - `name: "web_fetch"`
527
528 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
529
530 Permission policy for tool execution.
531
532 - `BetaManagedAgentsAlwaysAllowPolicy object`
533
534 Tool calls are automatically approved without user confirmation.
535
536 - `BetaManagedAgentsAlwaysAskPolicy object`
537
538 Tool calls require user confirmation before execution.
539
540 - `BetaManagedAgentsAutoPolicy object`
541
542 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
543
544 - `url_sources: BetaManagedAgentsWebFetchURLSources or null`
545
546 Which sources contribute URLs the tool may fetch, always in the object form. Null when not set, which allows every source.
547
548 - `client_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
549
550 Which custom tools' results contribute URLs that may be fetched. Null when not set, which allows every custom tool's results.
551
552 - `BetaManagedAgentsWebFetchURLSourceAll object`
553
554 Every URL from this source may be fetched. This is the default.
555
556 - `type: "all"`
557
558 - `BetaManagedAgentsWebFetchURLSourceNone object`
559
560 This source contributes no URLs that may be fetched.
561
562 - `type: "none"`
563
564 - `BetaManagedAgentsWebFetchURLSourceOnly object`
565
566 Only the named tools' results contribute URLs that may be fetched.
567
568 - `type: "only"`
569
570 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
571
572 The tools whose results contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "none" to allow no tool's results.
573
574 - `type: "tool_reference"`
575
576 Must be "tool_reference".
577
578 - `name: string`
579
580 Name of the tool. Compared exactly, so upper and lower case letters are different.
581
582 minLength: 1, maxLength: 128
583
584 - `BetaManagedAgentsWebFetchURLSourceExcept object`
585
586 Every tool's results contribute URLs that may be fetched, except the named tools' results.
587
588 - `type: "except"`
589
590 - `tools: array of BetaManagedAgentsWebFetchURLSourceToolReference`
591
592 The tools whose results do not contribute. Between 1 and 128 entries, each with a different name. An empty list is rejected; use "all" to leave out no tool's results.
593
594 - `type: "tool_reference"`
595
596 Must be "tool_reference".
597
598 - `name: string`
599
600 Name of the tool. Compared exactly, so upper and lower case letters are different.
601
602 minLength: 1, maxLength: 128
603
604 - `server_tool_results: BetaManagedAgentsWebFetchURLSourceToolFilter or null`
605
606 Which of the web_search and web_fetch tools' results contribute URLs that may be fetched. Null when not set, which allows both.
607
608 - `user_input: BetaManagedAgentsWebFetchURLSourceUserInput or null`
609
610 Whether URLs in the text of user messages may be fetched. Null when not set, which allows them.
611
612 - `BetaManagedAgentsWebFetchURLSourceAll object`
613
614 Every URL from this source may be fetched. This is the default.
615
616 - `BetaManagedAgentsWebFetchURLSourceNone object`
617
618 This source contributes no URLs that may be fetched.
619
620 - `allowed_domains: optional array of string`
621
622 - `blocked_domains: optional array of string`
623
624 - `max_content_tokens: optional number or null`
625
626 format: int32
627
628 - `BetaManagedAgentsWebSearchToolConfig object`
629
630 Configuration for the web_search tool.
631
632 - `type: "web_search"`
633
634 - `enabled: boolean`
635
636 - `name: "web_search"`
637
638 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
639
640 Permission policy for tool execution.
641
642 - `BetaManagedAgentsAlwaysAllowPolicy object`
643
644 Tool calls are automatically approved without user confirmation.
645
646 - `BetaManagedAgentsAlwaysAskPolicy object`
647
648 Tool calls require user confirmation before execution.
649
650 - `BetaManagedAgentsAutoPolicy object`
651
652 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
653
654 - `allowed_domains: optional array of string`
655
656 - `blocked_domains: optional array of string`
657
658 - `user_location: optional BetaManagedAgentsUserLocation or null`
659
660 Approximate user location for search result localization.
661
662 - `type: "approximate"`
663
664 Location precision. Only "approximate" is supported.
665
666 - `city: optional string or null`
667
668 City name.
669
670 minLength: 1, maxLength: 255
671
672 - `country: optional string or null`
673
674 Two-letter ISO 3166-1 country code, uppercase.
675
676 - `region: optional string or null`
677
678 Region or state name.
679
680 minLength: 1, maxLength: 255
681
682 - `timezone: optional string or null`
683
684 IANA timezone identifier, e.g. "America/Los_Angeles".
685
686 minLength: 1, maxLength: 255
687
688 - `default_config: BetaManagedAgentsAgentToolsetDefaultConfig`
689
690 Resolved default configuration for agent tools.
691
692 - `enabled: boolean`
693
694 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
695
696 Permission policy for tool execution.
697
698 - `BetaManagedAgentsAlwaysAllowPolicy object`
699
700 Tool calls are automatically approved without user confirmation.
701
702 - `BetaManagedAgentsAlwaysAskPolicy object`
703
704 Tool calls require user confirmation before execution.
705
706 - `BetaManagedAgentsAutoPolicy object`
707
708 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
709
710 - `BetaManagedAgentsMCPToolset object`
711
712 - `type: "mcp_toolset"`
713
714 - `configs: array of BetaManagedAgentsMCPToolConfig`
715
716 - `enabled: boolean`
758 - `mcp_server_name: string`
759
760 - `BetaManagedAgentsCustomTool object`
761
762 A custom tool as returned in API responses.
763
764 - `type: "custom"`
765
766 - `description: string`
767
768 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
769
770 JSON Schema for custom tool input parameters.
771
772 - `type: "object"`
773
774 - `properties: optional map[unknown] or null`
775
776 - `required: optional array of string or null`
717777
718778 - `name: string`
719779
720 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
721
722 Permission policy for tool execution.
723
724 - `BetaManagedAgentsAlwaysAllowPolicy object`
725
726 Tool calls are automatically approved without user confirmation.
727
728 - `BetaManagedAgentsAlwaysAskPolicy object`
729
730 Tool calls require user confirmation before execution.
731
732 - `BetaManagedAgentsAutoPolicy object`
733
734 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
735
736 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
737
738 Resolved default configuration for all tools from an MCP server.
739
740 - `enabled: boolean`
741
742 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
743
744 Permission policy for tool execution.
745
746 - `BetaManagedAgentsAlwaysAllowPolicy object`
747
748 Tool calls are automatically approved without user confirmation.
749
750 - `BetaManagedAgentsAlwaysAskPolicy object`
751
752 Tool calls require user confirmation before execution.
753
754 - `BetaManagedAgentsAutoPolicy object`
755
756 The server decides each tool call individually: it judges, from the tool, its input, and the session content so far, whether the call is safe to execute or high-risk, and evaluates it to allow when judged safe and to deny when judged high-risk. A call the server cannot reach a judgement on evaluates to ask.
757
758 - `mcp_server_name: string`
759
760 - `BetaManagedAgentsCustomTool object`
761
762 A custom tool as returned in API responses.
763
764 - `type: "custom"`
765
766 - `description: string`
767
768 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
769
770 JSON Schema for custom tool input parameters.
771
772 - `type: "object"`
773
774 - `properties: optional map[unknown] or null`
775
776 - `required: optional array of string or null`
780 - `version: number`
781
782 format: int32
783
784 - `BetaManagedAgentsAdvisor object`
785
786 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
787
788 - `type: "advisor"`
789
790 - `model: string`
791
792 The advisor model id.
793
794 - `BetaManagedAgentsSessionMultiagent20261001 object`
795
796 Resolved multiagent configuration with three members, as copied to the `session` at creation.
797
798 - `type: "multiagent_20261001"`
799
800 - `advisor: BetaManagedAgentsMultiagentAdvisor`
801
802 Whether the session's primary thread can consult an advisor model.
803
804 - `BetaManagedAgentsMultiagentAdvisorEnabled object`
805
806 The session's primary thread can consult `model` mid-turn.
807
808 - `type: "enabled"`
809
810 - `model: string`
811
812 The advisor model id.
813
814 - `BetaManagedAgentsMultiagentAdvisorDisabled object`
815
816 The agent has no advisor.
817
818 - `type: "disabled"`
819
820 - `subagents: BetaManagedAgentsSessionMultiagentSubagents`
821
822 Whether the agent can spawn session threads.
823
824 - `BetaManagedAgentsSessionMultiagentSubagentsEnabled object`
825
826 The agent can spawn session threads.
827
828 - `type: "enabled"`
829
830 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
831
832 Whether the agent can define inline agents, which are not saved, when it spawns session threads.
833
834 - `BetaManagedAgentsMultiagentInlineAgentsEnabled object`
835
836 The agent can define inline agents.
837
838 - `type: "enabled"`
839
840 - `BetaManagedAgentsMultiagentInlineAgentsDisabled object`
841
842 The agent cannot define inline agents.
843
844 - `type: "disabled"`
845
846 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
847
848 Full `agent` definitions of the predefined agents, which are saved agents that this agent can spawn as session threads.
849
850 - `type: "agent"`
851
852 - `id: string`
853
854 - `description: string or null`
855
856 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
857
858 - `model: BetaManagedAgentsModelConfig`
859
860 Model identifier and configuration.
777861
778862 - `name: string`
779863
780 - `version: number`
781
782 format: int32
783
784 - `BetaManagedAgentsAdvisor object`
785
786 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
787
788 - `type: "advisor"`
789
790 - `model: string`
791
792 The advisor model id.
864 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
865
866 - `system: string or null`
867
868 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
869
870 - `version: number`
871
872 format: int32
873
874 - `BetaManagedAgentsMultiagentSubagentsDisabled object`
875
876 The agent cannot spawn session threads.
877
878 - `type: "disabled"`
879
880 - `workflows: BetaManagedAgentsSessionMultiagentWorkflows`
881
882 Whether the agent can start workflow runs.
883
884 - `BetaManagedAgentsSessionMultiagentWorkflowsEnabled object`
885
886 The agent can start workflow runs.
887
888 - `type: "enabled"`
889
890 - `inline_agents: BetaManagedAgentsMultiagentInlineAgents`
891
892 Whether a run's plan can define inline agents, which are not saved.
893
894 - `predefined_agents: array of BetaManagedAgentsSessionThreadAgent`
895
896 Full `agent` definitions of the predefined agents, which are saved agents that a run's plan can use.
897
898 - `type: "agent"`
899
900 - `id: string`
901
902 - `description: string or null`
903
904 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
905
906 - `model: BetaManagedAgentsModelConfig`
907
908 Model identifier and configuration.
909
910 - `name: string`
911
912 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
913
914 - `system: string or null`
915
916 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
917
918 - `version: number`
919
920 format: int32
921
922 - `BetaManagedAgentsMultiagentWorkflowsDisabled object`
923
924 The agent cannot start workflow runs.
925
926 - `type: "disabled"`
793927
794928 - `name: string`
795929
796930
api/beta/sessions/threads Changed · +5457 / -3470 lines
### Beta Managed Agents Inline Agent
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
Nothing in the body moved in this read. What changed is above.
api/beta/sessions/threads/archive Changed · +53 / -6 lines
from line 143
143143
144144 Unique identifier for this thread.
145145
146 - `agent: BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor`
146 - `agent: BetaManagedAgentsSessionThreadAgent or BetaManagedAgentsAdvisor or BetaManagedAgentsInlineAgent`
147147
148148 Resolved agent definition for this thread. Snapshot of the agent at thread creation time.
149149
from line 221
221221
222222 - `"claude-haiku-4-5"`
223223
224 Fastest model with near-frontier intelligence
225
226224 - `"claude-haiku-4-5-20251001"`
227225
228 Fastest model with near-frontier intelligence
229
230226 - `"claude-opus-4-5"`
231227
232228 Powerful intelligence for long-running agents and coding
from line 763
767763
768764 The advisor model id.
769765
766 - `BetaManagedAgentsInlineAgent object`
767
768 An agent that has no Agent resource, and so no `id` or `version`. It is defined inline, in a workflow run's plan or when a session thread is spawned, and is not saved.
769
770 - `type: "inline"`
771
772 - `description: string or null`
773
774 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
775
776 - `type: "url"`
777
778 - `name: string`
779
780 - `url: string`
781
782 - `model: BetaManagedAgentsModelConfig`
783
784 Model identifier and configuration.
785
786 - `name: string`
787
788 The name that the agent's definition gave, or one that the server assigned.
789
790 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
791
792 - `BetaManagedAgentsAnthropicSkill object`
793
794 A resolved Anthropic-managed skill.
795
796 - `BetaManagedAgentsCustomSkill object`
797
798 A resolved user-created custom skill.
799
800 - `system: string or null`
801
802 - `tools: array of BetaManagedAgentsAgentToolset20260401 or BetaManagedAgentsMCPToolset or BetaManagedAgentsCustomTool`
803
804 - `BetaManagedAgentsAgentToolset20260401 object`
805
806 - `BetaManagedAgentsMCPToolset object`
807
808 - `BetaManagedAgentsCustomTool object`
809
810 A custom tool as returned in API responses.
811
770812 - `archived_at: string or null`
771813
772814 When the thread was archived. Null if not archived.
from line 941
899941
900942 format: int32
901943
944 - `workflow_run_id: string or null`
945
946 Identifier of the workflow run that created the thread, or `null` for any other thread.
947
902948## Example
903949
904950```bash
from line 1040
9941040 "web_fetch_requests": 0,
9951041 "web_search_requests": 3
9961042 }
997 }
1043 },
1044 "workflow_run_id": null
9981045}
9991046```
10001047