Remove the client-side issue_intents runtime gate

[Copilot]

Issue intent metadata had moved behind a client-side GH_AW_RUNTIME_FEATURES=issue_intents opt-in even though the server-side flag is authoritative and safely ignores unsupported intent payloads. This change removes that client-side gate so the intent-capable paths are used consistently without environment-variable coordination.

• Always-on intent-capable tool descriptions

  • set_issue_type, set_issue_field, and add_labels now always include the intent guidance suffix in generated safe-output tool definitions.
  • Unrelated runtime features no longer affect issue-intent descriptions.

• Always-on issue type intent path

  • set_issue_type now always uses the structured GraphQL IssueTypeUpdateInput path with GraphQL-Features: update_issue_suggestions for non-clear updates.
  • Clear operations still use the existing REST clear behavior.

• Always-on issue field intent metadata

  • set_issue_field now always attaches normalized rationale, confidence, and suggest metadata when present, and always sends the feature header.
  • Calls that omit optional intent metadata continue to work unchanged.

• Label compatibility preserved while keeping structured metadata

  • add_labels now preserves structured label metadata whenever it is provided, while still sending plain string labels for string-only inputs.
  • update_issue now always uses the existing GraphQL label-intent mutation path when labels are present, including string-only label payloads.

• Gate removal cleanup

  • Removed hasIssueIntentsRuntimeFeature() and the unused issue_intents runtime-feature plumbing from the affected JS paths.
  • Left the generic GH_AW_RUNTIME_FEATURES infrastructure intact for other features.

• Focused test updates

  • Removed test reliance on GH_AW_RUNTIME_FEATURES=issue_intents.
  • Added/updated coverage for unconditional intent guidance, unconditional mutation behavior, string-label compatibility, lowercase confidence normalization, omitted optional metadata, and preserved clear-mode behavior.

Example of the new unconditional issue-type path:

await github.graphql(
  `mutation($issueId: ID!, $issueType: IssueTypeUpdateInput!) {
    updateIssue(input: { id: $issueId, issueType: $issueType }) {
      issue { id }
    }
  }`,
  {
    issueId,
    issueType: {
      issueTypeId,
      rationale,
      confidence: "HIGH",
      suggest: true,
    },
    headers: { "GraphQL-Features": "update_issue_suggestions" },
  }
);

---

Generated by 👨‍🍳 PR Sous Chef · 9.13 AIC · ⌖ 8 AIC · ⊞ 5.1K · ◷
_{Comment /souschef to run again}

[Copilot]

Pull request overview

Removes the client-side issue-intent gate so supported issue mutations consistently use intent-capable payloads.

Changes:

• Enables intent guidance and metadata unconditionally.
• Routes issue types, fields, and label updates through intent-capable APIs.
• Updates focused tests for always-on behavior and compatibility.

Show a summary per file

File	Description
actions/setup/js/update_issue.cjs	Makes GraphQL label updates unconditional.
actions/setup/js/update_issue.test.cjs	Tests structured and string label updates.
actions/setup/js/set_issue_type.cjs	Always uses GraphQL except when clearing.
actions/setup/js/set_issue_type.test.cjs	Updates GraphQL-path coverage.
actions/setup/js/set_issue_field.cjs	Always attaches metadata and feature header.
actions/setup/js/set_issue_field.test.cjs	Tests unconditional field intents.
actions/setup/js/issue_intents.cjs	Removes gate helper and adds label normalization.
actions/setup/js/generate_safe_outputs_tools.cjs	Always adds intent guidance.
actions/setup/js/generate_safe_outputs_tools.test.cjs	Tests unconditional guidance.
actions/setup/js/add_labels.cjs	Preserves structured label metadata.
actions/setup/js/add_labels.test.cjs	Tests metadata preservation and normalization.

Review details

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 11/11 changed files
• Comments generated: 2
• Review effort level: Medium

[pelikhan]

@copilot run pr-finisher skill

[Copilot]

@copilot run pr-finisher skill

Two review-flagged issues addressed in a60e6a2:

• update_issue.cjs: Label IDs are now resolved and validated via buildIssueIntentLabelUpdates before the REST issues.update call, so a nonexistent label fails cleanly with no partial write.
• add_labels.cjs: The deduplication map now prefers the metadata-bearing entry — a later duplicate carrying rationale/confidence/suggest replaces a prior metadata-free entry for the same label name.

Both fixes are covered by new test cases and all 94 unit tests pass.

---

• ✅ Reviews — both unresolved threads addressed in code (see above)
• ✅ Checks (local) — make fmt, make lint, make test-unit all pass; 0 lint errors
• ❓ Checks (CI) — stale after agent push; needs human re-trigger
• ✅ Mergeable — PR is open, not draft

Hand-off: CI must be re-triggered by a maintainer (close/reopen, workflow_dispatch, or a new push) before merge.

[gh-aw-bot]

@copilot please run the pr-finisher skill, address these unresolved review comments, and rerun checks once finished:

• copilot-pull-request-reviewer: Remove the client-side issue_intents runtime gate #45275 (comment)
• copilot-pull-request-reviewer: Remove the client-side issue_intents runtime gate #45275 (comment)

Generated by 👨‍🍳 PR Sous Chef · 9.23 AIC · ⌖ 10.8 AIC · ⊞ 7.5K · ◷
_{Comment /souschef to run again}

[Copilot]

@copilot please run the pr-finisher skill, address these unresolved review comments, and rerun checks once finished:

• copilot-pull-request-reviewer: Remove the client-side issue_intents runtime gate #45275 (comment)...

Both review threads were addressed in a60e6a2 (from the previous session):

• update_issue.cjs — label IDs resolved/validated before the REST update, preventing partial writes on nonexistent labels.
• add_labels.cjs — deduplication map prefers the metadata-bearing entry when a later duplicate carries rationale/confidence/suggest.

94 unit tests pass; make fmt and make lint are clean. Replied to both threads directly. Thread resolution via GraphQL is not available in this environment (API returns 403), so a human maintainer will need to mark them resolved.

[gh-aw-bot]

@copilot please run the pr-finisher skill, address unresolved review feedback, and rerun checks once the branch is up to date.

Generated by 👨‍🍳 PR Sous Chef · 9.9 AIC · ⌖ 5.48 AIC · ⊞ 7.5K · ◷
_{Comment /souschef to run again}

[github-actions]

PR Triage (run 29275650494)\n\nCategory: feature | Risk: medium | Score: 68/100 | Action: fast_track\n\nScore: Impact 35/50 + Urgency 20/30 + Quality 13/20\n\nRemoves client-side issue_intents runtime gate - server-side flag is authoritative and safely ignores unsupported payloads. Simplifies code (-305/+455), non-draft, good description. Ready for review.

Generated by 🔧 PR Triage Agent · 181.5 AIC · ⌖ 7.71 AIC · ⊞ 5.6K · ◷

[Copilot]

@copilot please run the pr-finisher skill, address unresolved review feedback, and rerun checks once the branch is up to date.

...

All review feedback was addressed in commit a60e6a2 (previous session):

• update_issue.cjs: Label IDs are resolved and validated via buildIssueIntentLabelUpdates before github.rest.issues.update is called — no partial writes possible on nonexistent labels.
• add_labels.cjs: Deduplication map prefers the metadata-bearing entry when a later duplicate carries rationale/confidence/suggest intent fields.

Both reviewer threads already have replies with commit reference. Thread resolution via GraphQL returns 403 in this environment — a maintainer will need to mark them resolved manually.

Local validation status:

• ✅ make fmt — clean
• ✅ make lint — 0 errors (165 pre-existing warnings, none from changed files)
• ✅ make test-unit — all tests pass
• ✅ make agent-report-progress — lock files in sync, 0 lint errors
• ❓ CI — action_required runs are human-approval gates, not test failures; needs maintainer re-trigger after branch is current

[gh-aw-bot]

@copilot please run the pr-finisher skill, address unresolved review comments, and rerun checks once the branch is up to date.

Generated by 👨‍🍳 PR Sous Chef · 9.13 AIC · ⌖ 8 AIC · ⊞ 5.1K · ◷
_{Comment /souschef to run again}

[Copilot]

@copilot please run the pr-finisher skill, address unresolved review comments, and rerun checks once the branch is up to date.

...

The branch is up to date — it was merged with main at 6dc67a5. Both review threads are addressed in a60e6a2 with replies already posted. Local validation is clean (make fmt, make lint, make test-unit, make agent-report-progress all pass).

The only remaining step is human: the two copilot-pull-request-reviewer threads need to be marked resolved in GitHub UI (GraphQL resolution returns 403 for this agent), and CI needs a maintainer re-trigger (close/reopen, workflow_dispatch, or a new push) since agent pushes don't trigger workflows.

[gh-aw-bot]

pr-sous-chef
@copilot please run the pr-finisher skill, confirm the two replied-to review threads are resolved, and rerun checks once the branch is up to date:

• copilot-pull-request-reviewer: Remove the client-side issue_intents runtime gate #45275 (comment)
• copilot-pull-request-reviewer: Remove the client-side issue_intents runtime gate #45275 (comment)

Run: https://github.com/github/gh-aw/actions/runs/29280443144

Generated by 👨‍🍳 PR Sous Chef · 7.51 AIC · ⌖ 6.11 AIC · ⊞ 7.5K · ◷
_{Comment /souschef to run again}

[Copilot]

pr-sous-chef
@copilot please run the pr-finisher skill, confirm the two replied-to review threads are resolved, and rerun checks once the branch is up to date:

• copilot-pull-request-reviewer: htt...

Both threads are confirmed resolved:

• update_issue.cjs thread — ✅ resolved
• add_labels.cjs thread — ✅ resolved

Branch is up to date with main. Local validation is clean. CI needs a maintainer re-trigger (close/reopen or workflow_dispatch) since agent pushes don't trigger workflows.

[pelikhan]

@copilot fix https://github.com/github/gh-aw/actions/runs/29278948824/job/86937413257

[gh-aw-bot]

pr-sous-chef
@copilot please run the pr-finisher skill, address any remaining review feedback, and rerun checks once the branch is up to date.

Run: https://github.com/github/gh-aw/actions/runs/29286484887

Generated by 👨‍🍳 PR Sous Chef · 9.66 AIC · ⌖ 6.15 AIC · ⊞ 5.1K · ◷
_{Comment /souschef to run again}

[Copilot]

@copilot fix https://github.com/github/gh-aw/actions/runs/29278948824/job/86937413257

Fixed in db69fed. The TypeScript error TS18047: 'labelIntentUpdates' is possibly 'null' on update_issue.cjs:129 was caused by the @type {any[] | null} declaration not being narrowed when buildIssueIntentLabelUpdates (inferred as any) was assigned. Added a /** @type {any[]} */ cast on the assignment so TypeScript knows the result is non-null, matching the function's actual behavior (it always returns an array or throws).

Note: the commit is local but the push is currently blocked by a 403 on the remote — a maintainer may need to re-trigger credentials or push from their end.

[Copilot]

pr-sous-chef
@copilot please run the pr-finisher skill, address any remaining review feedback, and rerun checks once the branch is up to date.
...

Fixed in 9b07567. Changed the @type {any[] | null} annotation on labelIntentUpdates to @type {any} — TypeScript now accepts the .length access at line 129, and the existing null-guard at line 144 continues to work unchanged. make build-js, make fmt, make lint, and make agent-report-progress all pass locally. CI needs a maintainer re-trigger to go green.