DCF Model Rubric changedmanaged-agents/define-outcomes
Nearest release: v2.1.307, published an hour after this site recorded the change. Shown because the two are within 24 hours of each other. Nothing here says the release caused the edit.
Recorded here
Lines+9added
Lines−9removed
From line
23
where the diff opens
First seen
14 Aug 2026
this site's first read of the page
Recorded edits18to this page, all time
The whole hunk
from line 23, old and new numbered
/
from line 23
2323<Accordion title="Tips for writing effective rubrics">
2424 Structure the rubric as explicit, gradeable criteria, such as "The CSV contains a price column with numeric values" rather than "The data looks good." The grader scores each criterion independently, so vague criteria produce noisy evaluations.
2525
26 If you don't have a rubric on hand, try giving Claude an example of a known-good artifact and asking it to analyze what makes that content good, then turn that analysis into a rubric. This middle-ground approach often produces better results than writing criteria from scratch.
26 If you don't have a rubric, try giving Claude an example of a known-good artifact and asking it to analyze what makes that content good, then turn that analysis into a rubric. This middle-ground approach often produces better results than writing criteria with no example to start from.
2727</Accordion>
2828
2929Example rubric:
from line 285
285285
286286## Create a session with an outcome
287287
288The following examples create a [session](https://platform.claude.com/docs/en/managed-agents/sessions) for an existing [agent](https://platform.claude.com/docs/en/managed-agents/agent-setup) and [environment](https://platform.claude.com/docs/en/managed-agents/environments) (both created separately), then send a `user.define_outcome` event. The agent begins work immediately. No additional user message event is required.
288The following examples create a [session](https://platform.claude.com/docs/en/managed-agents/sessions) for an existing [agent](https://platform.claude.com/docs/en/managed-agents/agent-setup) and [environment](https://platform.claude.com/docs/en/managed-agents/environments) (both created separately), then send a `user.define_outcome` event. The agent starts work immediately. No additional user message event is required.
289289
290290<CodeGroup>
291291 ```bash cURL
from line 347
347347 title="Financial analysis on Costco",
348348 )
349349
350 # Define the outcome — agent starts working on receipt
350 # Define the outcome. The agent starts working on receipt.
351351 client.beta.sessions.events.send(
352352 session_id=session.id,
353353 events=[
from line 370
370370 title: "Financial analysis on Costco"
371371 });
372372
373 // Define the outcome — agent starts working on receipt
373 // Define the outcome. The agent starts working on receipt.
374374 await client.beta.sessions.events.send(session.id, {
375375 events: [
376376 {
from line 500
500500 title: "Financial analysis on Costco"
501501 )
502502
503 # Define the outcome — agent starts working on receipt
503 # Define the outcome. The agent starts working on receipt.
504504 client.beta.sessions.events.send_(
505505 session.id,
506506 events: [
from line 527
527527* `agent.*` events (such as messages and tool use) show progress toward the outcome.
528528* `span.outcome_evaluation_*` events are only emitted for outcome-oriented sessions and show the number of iteration loops and the grader's feedback process.
529529* You can also send `user.message` [events](https://platform.claude.com/docs/en/managed-agents/reference#event-types) to an outcome-oriented session to direct the agent's work as it progresses, but it isn't required: the agent works toward the outcome on its own, iterating until it succeeds or runs out of iterations.
530* A `user.interrupt` event pauses work on the current outcome and marks the `result` (csharp, go: `Result`) of the `span.outcome_evaluation_end` event as `interrupted`, allowing you to kick off a new outcome.
530* A `user.interrupt` event pauses work on the current outcome and marks the `result` (csharp, go: `Result`) of the `span.outcome_evaluation_end` event as `interrupted`, allowing you to start a new outcome.
531531* After the final outcome evaluation, the session can be continued as a conversational session, or a new outcome can be started. The session retains history of the prior outcome.
532532
533533### Define outcome user event
534534
535535<Note>
536 Only one outcome is supported at a time, but you may chain outcomes in sequence. To do this, send a new `user.define_outcome` event after the terminal `span.outcome_evaluation_end` event of the previous outcome.
536 Only one outcome is supported at a time, but you can chain outcomes in sequence. To do this, send a new `user.define_outcome` event after the terminal `span.outcome_evaluation_end` event of the previous outcome.
537537</Note>
538538
539539This is the event you send to initiate an outcome. It is echoed back on receipt, including a `processed_at` (java, php: `processedAt`; csharp, go: `ProcessedAt`) timestamp and `outcome_id` (java: `outcomeId`; php: `outcomeID`; csharp, go: `OutcomeID`).
from line 577
577577
578578### Outcome evaluation end
579579
580Emitted when an outcome evaluation cycle ends: after the grader finishes evaluating one iteration, or when the session is interrupted while an outcome is active. The `result` field indicates what happens next.
580Emitted when an outcome evaluation cycle ends: after the grader finishes evaluating one iteration, or when the session is interrupted while an outcome is active. The `result` field indicates what occurs next.
581581
582582| Result | Next |
583583| ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
from line 863
863863 </Card>
864864
865865 <Card title="Session event stream" icon="lightning" href="https://platform.claude.com/docs/en/managed-agents/events-and-streaming">
866 Send events, stream responses, and interrupt or redirect your session mid-execution.
866 Send events, stream responses, and interrupt or redirect your session while it runs.
867867 </Card>
868868
869869 <Card title="Adding files" icon="file" href="https://platform.claude.com/docs/en/managed-agents/files">
No line in this hunk matches that.