The whole hunk
from line 1, old and new numbered
/
lines
from line 1
1---
2title: Create Agent
3url: https://platform.claude.com/docs/en/api/beta/agents/create
4---
5
16# Create Agent
27
38**POST** `/v1/agents`
from line 17
1217
1318 - `string`
1419
15 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 41 more`
20 - `"message-batches-2024-09-24" or "prompt-caching-2024-07-31" or "computer-use-2024-10-22" or 42 more`
1621
1722 - `"message-batches-2024-09-24"`
1823
from line 65
6065
6166 - `"user-profiles-2026-08-18"`
6267
68 - `"user-profiles-2026-09-04"`
69
6370 - `"advisor-tool-2026-03-01"`
6471
6572 - `"managed-agents-2026-04-01"`
from line 109
102109
103110 - `"mid-conversation-system-clear-at-2026-08-21"`
104111
112- `"anthropic-workspace-id": optional string`
113
105114## Body parameters
106115
107116- `model: BetaManagedAgentsModel or BetaManagedAgentsModelConfigParams`
from line 273
264273
265274 MCP servers this agent connects to. Maximum 20. Names must be unique within the array. Every server must be referenced by an `mcp_toolset` in `tools`; unreferenced servers are rejected. See the [MCP connector guide](https://platform.claude.com/docs/en/managed-agents/mcp-connector).
266275
276 - `type: "url"`
277
267278 - `name: string`
268279
269280 Unique name for this server, referenced by mcp_toolset configurations. 1-255 characters.
from line 281
270281
271282 minLength: 1, maxLength: 255
272283
273 - `type: "url"`
274
275284 - `url: string`
276285
277286 Endpoint URL for the MCP server.
from line 295
286295
287296 A coordinator topology: the session's primary thread orchestrates work by spawning session threads, each running an agent drawn from the `agents` roster.
288297
298 - `type: "coordinator"`
299
289300 - `agents: array of BetaManagedAgentsMultiagentRosterEntryParams`
290301
291302 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).
from line 307
296307
297308 Specification for an Agent. Provide a specific `version` or use the short-form `agent="agent_id"` for the most recent version
298309
310 - `type: "agent"`
311
299312 - `id: string`
300313
301314 The `agent` ID.
from line 315
302315
303316 minLength: 1, maxLength: 128
304317
305 - `type: "agent"`
306
307318 - `version: optional number`
308319
309320 The specific `agent` version to use. Omit to use the latest version. Must be at least 1 if specified.
from line 331
320331
321332 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`.
322333
334 - `type: "advisor"`
335
323336 - `model: string`
324337
325338 A Claude model id. The model must be permitted as an advisor for this agent's model — see the sessions/threads/advisor spec.
from line 339
326339
327340 minLength: 1, maxLength: 256
328341
329 - `type: "advisor"`
330
331 - `type: "coordinator"`
332
333342- `skills: optional array of BetaManagedAgentsSkillParams`
334343
335344 Skills available to the agent.
from line 347
338347
339348 An Anthropic-managed skill.
340349
350 - `type: "anthropic"`
351
341352 - `skill_id: string`
342353
343354 Identifier of the Anthropic skill (e.g., "xlsx").
from line 355
344355
345356 minLength: 1, maxLength: 64
346357
347 - `type: "anthropic"`
348
349358 - `version: optional string or null`
350359
351360 Version to pin. Defaults to latest if omitted.
from line 365
356365
357366 A user-created custom skill.
358367
368 - `type: "custom"`
369
359370 - `skill_id: string`
360371
361372 Tagged ID of the custom skill (e.g., "skill_01XJ5...").
from line 373
362373
363374 minLength: 1, maxLength: 64
364375
365 - `type: "custom"`
366
367376 - `version: optional string or null`
368377
369378 Version to pin. Defaults to latest if omitted.
from line 403
394403
395404 Configuration override for the bash tool.
396405
406 - `type: optional "bash"`
407
397408 - `name: "bash"`
398409
399410 Must be "bash".
from line 413
402413
403414 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
404415
405 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
416 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
406417
407418 Permission policy for tool execution.
408419
from line 429
418429
419430 - `type: "always_ask"`
420431
421 - `type: optional "bash"`
432 - `BetaManagedAgentsAutoPolicy object`
422433
434 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.
435
436 - `type: "auto"`
437
423438 - `BetaManagedAgentsEditToolConfigParams object`
424439
425440 Configuration override for the edit tool.
426441
442 - `type: optional "edit"`
443
427444 - `name: "edit"`
428445
429446 Must be "edit".
from line 449
432449
433450 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
434451
435 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
452 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
436453
437454 Permission policy for tool execution.
438455
from line 461
444461
445462 Tool calls require user confirmation before execution.
446463
447 - `type: optional "edit"`
464 - `BetaManagedAgentsAutoPolicy object`
448465
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
449468 - `BetaManagedAgentsReadToolConfigParams object`
450469
451470 Configuration override for the read tool.
452471
472 - `type: optional "read"`
473
453474 - `name: "read"`
454475
455476 Must be "read".
from line 479
458479
459480 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
460481
461 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
482 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
462483
463484 Permission policy for tool execution.
464485
from line 491
470491
471492 Tool calls require user confirmation before execution.
472493
473 - `type: optional "read"`
494 - `BetaManagedAgentsAutoPolicy object`
474495
496 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.
497
475498 - `BetaManagedAgentsWriteToolConfigParams object`
476499
477500 Configuration override for the write tool.
478501
502 - `type: optional "write"`
503
479504 - `name: "write"`
480505
481506 Must be "write".
from line 509
484509
485510 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
486511
487 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
512 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
488513
489514 Permission policy for tool execution.
490515
from line 521
496521
497522 Tool calls require user confirmation before execution.
498523
499 - `type: optional "write"`
524 - `BetaManagedAgentsAutoPolicy object`
500525
526 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.
527
501528 - `BetaManagedAgentsGlobToolConfigParams object`
502529
503530 Configuration override for the glob tool.
504531
532 - `type: optional "glob"`
533
505534 - `name: "glob"`
506535
507536 Must be "glob".
from line 539
510539
511540 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
512541
513 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
542 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
514543
515544 Permission policy for tool execution.
516545
from line 551
522551
523552 Tool calls require user confirmation before execution.
524553
525 - `type: optional "glob"`
554 - `BetaManagedAgentsAutoPolicy object`
526555
556 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.
557
527558 - `BetaManagedAgentsGrepToolConfigParams object`
528559
529560 Configuration override for the grep tool.
530561
562 - `type: optional "grep"`
563
531564 - `name: "grep"`
532565
533566 Must be "grep".
from line 569
536569
537570 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
538571
539 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
572 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
540573
541574 Permission policy for tool execution.
542575
from line 581
548581
549582 Tool calls require user confirmation before execution.
550583
551 - `type: optional "grep"`
584 - `BetaManagedAgentsAutoPolicy object`
552585
586 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.
587
553588 - `BetaManagedAgentsWebFetchToolConfigParams object`
554589
555590 Configuration override for the web_fetch tool.
556591
592 - `type: optional "web_fetch"`
593
557594 - `name: "web_fetch"`
558595
559596 Must be "web_fetch".
from line 613
576613
577614 format: int32
578615
579 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
616 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
580617
581618 Permission policy for tool execution.
582619
from line 625
588625
589626 Tool calls require user confirmation before execution.
590627
591 - `type: optional "web_fetch"`
628 - `BetaManagedAgentsAutoPolicy object`
592629
630 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.
631
593632 - `BetaManagedAgentsWebSearchToolConfigParams object`
594633
595634 Configuration override for the web_search tool.
596635
636 - `type: optional "web_search"`
637
597638 - `name: "web_search"`
598639
599640 Must be "web_search".
from line 651
610651
611652 Whether this tool is enabled and available to Claude. Overrides the default_config setting.
612653
613 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
654 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
614655
615656 Permission policy for tool execution.
616657
from line 663
622663
623664 Tool calls require user confirmation before execution.
624665
625 - `type: optional "web_search"`
666 - `BetaManagedAgentsAutoPolicy object`
626667
668 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.
669
627670 - `user_location: optional BetaManagedAgentsUserLocation or null`
628671
629672 Approximate user location for search result localization.
from line 705
662705
663706 Whether tools are enabled and available to Claude by default. Defaults to true if not specified.
664707
665 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
708 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
666709
667710 Permission policy for tool execution.
668711
from line 717
674717
675718 Tool calls require user confirmation before execution.
676719
720 - `BetaManagedAgentsAutoPolicy object`
721
722 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.
723
677724 - `BetaManagedAgentsMCPToolsetParams object`
678725
679726 Configuration for tools from an MCP server defined in `mcp_servers`.
680727
728 - `type: "mcp_toolset"`
729
681730 - `mcp_server_name: string`
682731
683732 Name of the MCP server. Must match a server name from the mcp_servers array. 1-255 characters.
from line 733
684733
685734 minLength: 1, maxLength: 255
686735
687 - `type: "mcp_toolset"`
688
689736 - `configs: optional array of BetaManagedAgentsMCPToolConfigParams`
690737
691738 Per-tool configuration overrides.
from line 747
700747
701748 Whether this tool is enabled. Overrides the `default_config` setting.
702749
703 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
750 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
704751
705752 Permission policy for tool execution.
706753
from line 759
712759
713760 Tool calls require user confirmation before execution.
714761
762 - `BetaManagedAgentsAutoPolicy object`
763
764 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.
765
715766 - `default_config: optional BetaManagedAgentsMCPToolsetDefaultConfigParams or null`
716767
717768 Default configuration for all tools from an MCP server.
from line 771
720771
721772 Whether tools are enabled by default. Defaults to true if not specified.
722773
723 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or null`
774 - `permission_policy: optional BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy or null`
724775
725776 Permission policy for tool execution.
726777
from line 783
732783
733784 Tool calls require user confirmation before execution.
734785
786 - `BetaManagedAgentsAutoPolicy object`
787
788 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.
789
735790 - `BetaManagedAgentsCustomToolParams object`
736791
737792 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.
738793
794 - `type: "custom"`
795
739796 - `description: string`
740797
741798 Description of what the tool does, shown to the agent to help it decide when to use the tool.
from line 815
758815
759816 minLength: 1, maxLength: 128
760817
761 - `type: "custom"`
762
763818## Returns
764819
765820- `BetaManagedAgentsAgent object`
from line 821
766821
767822 A Managed Agents `agent`.
768823
824 - `type: "agent"`
825
769826 - `id: string`
770827
771828 - `archived_at: string or null`
from line 841
784841
785842 - `mcp_servers: array of BetaManagedAgentsMCPServerURLDefinition`
786843
787 - `name: string`
788
789844 - `type: "url"`
790845
846 - `name: string`
847
791848 - `url: string`
792849
793850 - `metadata: map[string]`
from line 973
916973
917974 Resolved coordinator topology with a concrete agent roster.
918975
976 - `type: "coordinator"`
977
919978 - `agents: array of BetaManagedAgentsAgentReference or BetaManagedAgentsAdvisor`
920979
921980 Agents the coordinator may spawn as session threads, each resolved to a specific version.
from line 983
924983
925984 A resolved agent reference with a concrete version.
926985
927 - `id: string`
928
929986 - `type: "agent"`
930987
988 - `id: string`
989
931990 - `version: number`
932991
933992 format: int32
from line 995
936995
937996 Platform advisor roster entry: a model the session's primary thread may consult mid-turn.
938997
998 - `type: "advisor"`
999
9391000 - `model: string`
9401001
9411002 The advisor model id.
9421003
943 - `type: "advisor"`
944
945 - `type: "coordinator"`
946
9471004 - `name: string`
9481005
9491006 - `skills: array of BetaManagedAgentsAnthropicSkill or BetaManagedAgentsCustomSkill`
from line 1009
9521009
9531010 A resolved Anthropic-managed skill.
9541011
955 - `skill_id: string`
956
9571012 - `type: "anthropic"`
9581013
1014 - `skill_id: string`
1015
9591016 - `version: string`
9601017
9611018 - `BetaManagedAgentsCustomSkill object`
from line 1019
9621019
9631020 A resolved user-created custom skill.
9641021
965 - `skill_id: string`
966
9671022 - `type: "custom"`
9681023
1024 - `skill_id: string`
1025
9691026 - `version: string`
9701027
9711028 - `system: string or null`
from line 1031
9741031
9751032 - `BetaManagedAgentsAgentToolset20260401 object`
9761033
1034 - `type: "agent_toolset_20260401"`
1035
9771036 - `configs: array of BetaManagedAgentsAgentToolConfig`
9781037
9791038 - `BetaManagedAgentsBashToolConfig object`
from line 1039
9801039
9811040 Configuration for the bash tool.
9821041
1042 - `type: "bash"`
1043
9831044 - `enabled: boolean`
9841045
9851046 - `name: "bash"`
9861047
987 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1048 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
9881049
9891050 Permission policy for tool execution.
9901051
from line 1061
10001061
10011062 - `type: "always_ask"`
10021063
1003 - `type: "bash"`
1064 - `BetaManagedAgentsAutoPolicy object`
10041065
1066 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.
1067
1068 - `type: "auto"`
1069
10051070 - `BetaManagedAgentsEditToolConfig object`
10061071
10071072 Configuration for the edit tool.
10081073
1074 - `type: "edit"`
1075
10091076 - `enabled: boolean`
10101077
10111078 - `name: "edit"`
10121079
1013 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1080 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
10141081
10151082 Permission policy for tool execution.
10161083
from line 1089
10221089
10231090 Tool calls require user confirmation before execution.
10241091
1025 - `type: "edit"`
1092 - `BetaManagedAgentsAutoPolicy object`
10261093
1094 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.
1095
10271096 - `BetaManagedAgentsReadToolConfig object`
10281097
10291098 Configuration for the read tool.
10301099
1100 - `type: "read"`
1101
10311102 - `enabled: boolean`
10321103
10331104 - `name: "read"`
10341105
1035 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1106 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
10361107
10371108 Permission policy for tool execution.
10381109
from line 1115
10441115
10451116 Tool calls require user confirmation before execution.
10461117
1047 - `type: "read"`
1118 - `BetaManagedAgentsAutoPolicy object`
10481119
1120 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.
1121
10491122 - `BetaManagedAgentsWriteToolConfig object`
10501123
10511124 Configuration for the write tool.
10521125
1126 - `type: "write"`
1127
10531128 - `enabled: boolean`
10541129
10551130 - `name: "write"`
10561131
1057 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1132 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
10581133
10591134 Permission policy for tool execution.
10601135
from line 1141
10661141
10671142 Tool calls require user confirmation before execution.
10681143
1069 - `type: "write"`
1144 - `BetaManagedAgentsAutoPolicy object`
10701145
1146 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.
1147
10711148 - `BetaManagedAgentsGlobToolConfig object`
10721149
10731150 Configuration for the glob tool.
10741151
1152 - `type: "glob"`
1153
10751154 - `enabled: boolean`
10761155
10771156 - `name: "glob"`
10781157
1079 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1158 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
10801159
10811160 Permission policy for tool execution.
10821161
from line 1167
10881167
10891168 Tool calls require user confirmation before execution.
10901169
1091 - `type: "glob"`
1170 - `BetaManagedAgentsAutoPolicy object`
10921171
1172 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.
1173
10931174 - `BetaManagedAgentsGrepToolConfig object`
10941175
10951176 Configuration for the grep tool.
10961177
1178 - `type: "grep"`
1179
10971180 - `enabled: boolean`
10981181
10991182 - `name: "grep"`
11001183
1101 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1184 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
11021185
11031186 Permission policy for tool execution.
11041187
from line 1193
11101193
11111194 Tool calls require user confirmation before execution.
11121195
1113 - `type: "grep"`
1196 - `BetaManagedAgentsAutoPolicy object`
11141197
1198 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.
1199
11151200 - `BetaManagedAgentsWebFetchToolConfig object`
11161201
11171202 Configuration for the web_fetch tool.
11181203
1204 - `type: "web_fetch"`
1205
11191206 - `enabled: boolean`
11201207
11211208 - `name: "web_fetch"`
11221209
1123 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1210 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
11241211
11251212 Permission policy for tool execution.
11261213
from line 1219
11321219
11331220 Tool calls require user confirmation before execution.
11341221
1135 - `type: "web_fetch"`
1222 - `BetaManagedAgentsAutoPolicy object`
11361223
1224 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.
1225
11371226 - `allowed_domains: optional array of string`
11381227
11391228 - `blocked_domains: optional array of string`
from line 1235
11461235
11471236 Configuration for the web_search tool.
11481237
1238 - `type: "web_search"`
1239
11491240 - `enabled: boolean`
11501241
11511242 - `name: "web_search"`
11521243
1153 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1244 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
11541245
11551246 Permission policy for tool execution.
11561247
from line 1253
11621253
11631254 Tool calls require user confirmation before execution.
11641255
1165 - `type: "web_search"`
1256 - `BetaManagedAgentsAutoPolicy object`
11661257
1258 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.
1259
11671260 - `allowed_domains: optional array of string`
11681261
11691262 - `blocked_domains: optional array of string`
from line 1297
12041297
12051298 - `enabled: boolean`
12061299
1207 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1300 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
12081301
12091302 Permission policy for tool execution.
12101303
from line 1309
12161309
12171310 Tool calls require user confirmation before execution.
12181311
1219 - `type: "agent_toolset_20260401"`
1312 - `BetaManagedAgentsAutoPolicy object`
12201313
1314 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.
1315
12211316 - `BetaManagedAgentsMCPToolset object`
12221317
1318 - `type: "mcp_toolset"`
1319
12231320 - `configs: array of BetaManagedAgentsMCPToolConfig`
12241321
12251322 - `enabled: boolean`
from line 1323
12261323
12271324 - `name: string`
12281325
1229 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1326 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
12301327
12311328 Permission policy for tool execution.
12321329
from line 1335
12381335
12391336 Tool calls require user confirmation before execution.
12401337
1338 - `BetaManagedAgentsAutoPolicy object`
1339
1340 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.
1341
12411342 - `default_config: BetaManagedAgentsMCPToolsetDefaultConfig`
12421343
12431344 Resolved default configuration for all tools from an MCP server.
from line 1345
12441345
12451346 - `enabled: boolean`
12461347
1247 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy`
1348 - `permission_policy: BetaManagedAgentsAlwaysAllowPolicy or BetaManagedAgentsAlwaysAskPolicy or BetaManagedAgentsAutoPolicy`
12481349
12491350 Permission policy for tool execution.
12501351
from line 1357
12561357
12571358 Tool calls require user confirmation before execution.
12581359
1259 - `mcp_server_name: string`
1360 - `BetaManagedAgentsAutoPolicy object`
12601361
1261 - `type: "mcp_toolset"`
1362 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.
12621363
1364 - `mcp_server_name: string`
1365
12631366 - `BetaManagedAgentsCustomTool object`
12641367
12651368 A custom tool as returned in API responses.
12661369
1370 - `type: "custom"`
1371
12671372 - `description: string`
12681373
12691374 - `input_schema: BetaManagedAgentsCustomToolInputSchema`
from line 1382
12771382 - `required: optional array of string or null`
12781383
12791384 - `name: string`
1280
1281 - `type: "custom"`
1282
1283 - `type: "agent"`
12841385
12851386 - `updated_at: string`
12861387