Add reasoning effort support to assign-to-agent

[Copilot]

assign-to-agent can now forward optional, model-specific reasoning effort through the existing agent assignment API. Unsupported agents, models, or values emit a warning and continue without the field.

• Configuration

  • Adds templatable reasoning_effort string support.
  • Preserves expressions until runtime.
  • Updates schemas, fixtures, references, and changeset.

• Runtime

  • Sends supported values as agent_assignment.reasoning_effort.
  • Centralizes agent/model capability validation.
  • Includes reasoning effort in staged previews.
  • Preserves existing issue and pull-request assignment behavior.

• Coverage

  • Covers literals, expressions, omitted and invalid values, model-specific values, fallback behavior, and issue/PR targets.

safe-outputs:
  assign-to-agent:
    name: copilot
    model: o3
    reasoning_effort: high

Dynamic values are also supported:

reasoning_effort: ${{ inputs.reasoning_effort }}

• Fixes Add reasoning_effort support to assign-to-agent safe output #60332

[Copilot]

🟡 Changes recommended

Dynamic special-character values can break handler configuration parsing, while explicit empty values bypass the required warning.

Get a fresh assessment by requesting another Copilot review.

Pull request overview

Adds optional model-specific reasoning effort to assign-to-agent, spanning configuration, runtime validation, previews, tests, fixtures, and documentation.

Changes:

• Propagates reasoning_effort into compatible assignment requests.
• Adds capability-aware validation with warning-based fallback.
• Updates coverage, examples, schema, and release notes.

File summaries

File	Description
scratchpad/safe-outputs-specification.md	Documents payload behavior.
pkg/workflow/safe_outputs_handler_registry_assignments.go	Propagates the configuration.
pkg/workflow/assign_to_agent.go	Defines the new field.
pkg/workflow/assign_to_agent_test.go	Tests parsing and types.
pkg/parser/schemas/main_workflow_schema.json	Adds schema support.
pkg/cli/workflows/test-copilot-assign-to-agent.md	Adds a supported literal fixture.
pkg/cli/workflows/test-assign-to-agent-with-model.md	Adds an unsupported-model fixture.
pkg/cli/workflows/test-assign-to-agent-with-dynamic-reasoning-effort.md	Adds an expression fixture.
docs/src/content/docs/reference/safe-outputs.md	Documents safe-output usage.
docs/src/content/docs/reference/frontmatter-full.md	Updates generated reference content.
docs/src/content/docs/reference/copilot-cloud-agent.mdx	Documents Copilot configuration.
actions/setup/js/assign_to_agent.test.cjs	Tests previews and issue/PR propagation.
actions/setup/js/assign_to_agent.cjs	Passes configuration into assignments.
actions/setup/js/assign_agent_helpers.test.cjs	Tests capability validation.
actions/setup/js/assign_agent_helpers.cjs	Validates and builds the API payload.
.changeset/patch-assign-to-agent-reasoning-effort.md	Records the feature change.

Review details

• Files reviewed: 16/16 changed files
• Comments generated: 2
• Review effort level: Balanced

---

💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.

[github-actions]

✅ Test Quality Sentinel completed test quality analysis.

Test Quality Sentinel skipped because pre-fetch PR data was unavailable: unable to fetch test file diff

🧪 Test quality analysis by Test Quality Sentinel

[github-actions]

⚠️ Security scanning failed for Design Decision Gate 🏗️. Review the logs for details.

No ADR enforcement needed: PR does not have the 'implementation' label and has <=100 new lines of code in business logic directories.

🏗️ ADR gate enforced by Design Decision Gate 🏗️

[github-actions]

🧠 Matt Pocock Skills Reviewer failed during the skills-based review.

Warning

Threat Detection Engine Failure — The analysis engine could not complete. This is a tooling failure, not a security finding.

What happened

The threat detection engine failed to produce results.

Review the workflow run logs for details.

🧠 Reviewed using Matt Pocock's skills by Matt Pocock Skills Reviewer

[github-actions]

✅ Ponytail Reviewer completed successfully!

Warning

Threat Detection Engine Failure — The analysis engine could not complete. This is a tooling failure, not a security finding.

What happened

The threat detection engine failed to produce results.

Review the workflow run logs for details.

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• ab.chatgpt.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "ab.chatgpt.com"

See Network Configuration for more information.

Generated by Ponytail Reviewer for #60346

[github-actions]

⚠️ PR Code Quality Reviewer failed during code quality review.

Warning

Threat Detection Engine Failure — The analysis engine could not complete. This is a tooling failure, not a security finding.

What happened

The threat detection engine failed to produce results.

Review the workflow run logs for details.

🔎 Code quality review by PR Code Quality Reviewer

[pelikhan]

@copilot this is a lost cause, there will be more models in the future with different reasoning model. Do not try to enforce besides the enum values.

[Copilot]

Implemented in f8fda4e. reasoning_effort is now validated only against the enum (none, minimal, low, medium, high, xhigh) and forwarded without agent/model capability checks; tests cover an unknown future model.

[pelikhan]

@copilot use - instead of _ like other fields in the safe outputs from matter.

[Copilot]

Implemented in c07e8f4. The safe-output frontmatter and handler configuration now use reasoning-effort consistently; only the required REST API payload remains agent_assignment.reasoning_effort.

[github-actions]

🎉 This pull request is included in a new release.

Release: v0.89.8