Fix --use-samples replay for call/dispatch workflow tools and enable replace-label sample schema

[Copilot]

--use-samples could not deterministically replay call-workflow and dispatch-workflow because replay emitted family tool names (call_workflow / dispatch_workflow) instead of the generated workflow-specific tool names. replace-label also could not accept samples at compile time because its MCP tool schema was missing from safe_outputs_tools.json.

• Replay tool-name resolution for dynamic workflow handlers

  • Updated sample entry generation to map workflow_name → normalized dynamic tool name (e.g., worker-a → worker_a) for call-workflow and dispatch-workflow.
  • Preserved argument shaping for dynamic calls:
    • inputs payload is passed as tool args
    • dispatch-workflow also carries ref when provided

• Schema coverage for replace_label samples

  • Added replace_label tool schema to runtime and embedded tool schema catalogs:
    • actions/setup/js/safe_outputs_tools.json
    • pkg/workflow/js/safe_outputs_tools.json
  • This aligns compile-time sample validation with runtime handler availability.

• Regression-focused tests

  • Added coverage that collectSampleEntries emits workflow-specific tool names for dynamic workflow samples.
  • Extended dynamic-sample validation coverage to include replace-label sample configuration.

// dynamic workflow sample now targets generated tool name, not "call_workflow"
{
  "tool": "worker_a",
  "arguments": { "sentinel": "sample" }
}

• Fixes samples mode cannot replay call-workflow, dispatch-workflow, or replace-label #54811

[github-actions]

❌ Test Quality Sentinel failed during test quality analysis.

🧪 Test quality analysis by Test Quality Sentinel

[github-actions]

❌ Design Decision Gate 🏗️ failed during design decision gate check.

🏗️ ADR gate enforced by Design Decision Gate 🏗️

[github-actions]

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

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

[github-actions]

❌ Ponytail Reviewer failed. Please review the logs for details.

Generated by Ponytail Reviewer for #54813

[github-actions]

⚠️ PR Code Quality Reviewer failed during code quality review.

🔎 Code quality review by PR Code Quality Reviewer

[Copilot]

Pull request overview

Fixes #54811 by making sample replay target workflow-specific MCP tools and enabling replace-label sample validation.

Changes:

• Normalizes dynamic workflow sample tool names and reshapes arguments.
• Adds replace_label to both schema catalogs.
• Adds regression coverage for replay and validation.

Show a summary per file

File	Description
pkg/workflow/samples_replay.go	Resolves dynamic workflow sample tool names and arguments.
pkg/workflow/samples_validation_test.go	Tests dynamic replay and replace-label validation.
pkg/workflow/js/safe_outputs_tools.json	Adds the embedded replace_label schema.
actions/setup/js/safe_outputs_tools.json	Adds the runtime replace_label schema.

Review details

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

• Files reviewed: 4/4 changed files
• Comments generated: 0
• Review effort level: Balanced

[gh-aw-bot]

@copilot this PR looks close, but it still needs a maintainer-ready finish.

• Please review the unresolved feedback/check status on the current HEAD, address any remaining blockers, refresh the branch if needed, and leave a concise maintainer handoff with commands and outcomes.
• Then run the pr-finisher skill before handing back.

Run: https://github.com/github/gh-aw/actions/runs/32591161200

Generated by 👨‍🍳 PR Sous Chef · gpt54 · 8.55 AIC · ⌖ 8.26 AIC · ⊞ 9.5K · ◷
_{Comment /souschef to run again}

[Copilot]

@copilot this PR looks close, but it still needs a maintainer-ready finish.

• Please review the unresolved feedback/check status on the current HEAD, address any remaining blockers, refresh the branc...

Addressed in f4c1d64:

• Fixed the real blocker from lint-go-golangci (incremental lint on HEAD a335750): mapsloop modernize violation in pkg/workflow/samples_replay.go:86 — replaced the manual copy loop with maps.Copy.
• Fixed the real blocker from JS Tests (shard 4/4): safe_outputs_mcp_schema_validation.test.cjs failed because replace_label's label_to_add/label_to_remove descriptions were too short (47 chars) and lacked example/format guidance. Extended both descriptions with concrete examples (e.g. 'triage', 'in-progress') in both actions/setup/js/safe_outputs_tools.json and pkg/workflow/js/safe_outputs_tools.json.
• The other failing bot checks (Test Quality Sentinel, Design Decision Gate, Matt Pocock Skills Reviewer, Ponytail Reviewer, PR Code Quality Reviewer) had no failed jobs in their runs — they're commentary-style agentic workflows, not gating CI failures, and had no unresolved review threads to action.

Verified locally on the new commit: go build ./..., golangci-lint run ./pkg/workflow/... (0 issues), go test ./pkg/workflow/... (pass), and the full JS suite in actions/setup/js (only pre-existing unrelated flake in claude_harness.test.cjs, passes in isolation).

No open review threads on this PR. CI on this new HEAD needs a maintainer re-trigger since my push doesn't re-run workflows.

[gh-aw-bot]

@copilot this PR still needs forward progress.

• This branch is in a conflicting/dirty merge state. Please run make merge-main, resolve the conflicts cleanly, rerun the relevant validation, and push the result.
• After that, run the pr-finisher skill and leave a concise maintainer handoff with commands and outcomes.

Run: https://github.com/github/gh-aw/actions/runs/32595245207

Generated by 👨‍🍳 PR Sous Chef · gpt54 · 13 AIC · ⌖ 9.48 AIC · ⊞ 9.5K · ◷
_{Comment /souschef to run again}

[Copilot]

@copilot this PR still needs forward progress.

• This branch is in a conflicting/dirty merge state. Please run make merge-main, resolve the conflicts cleanly, rerun the relevant validation, and pus...

Merged in f3df3fd. Fetched origin/main and ran git merge origin/main; the only real conflict was in pkg/workflow/samples_replay.go's import block — this branch added maps while an unrelated concurrent fix on main (#54814) added sort. Kept both imports, then ran make build, make recompile (289/289 workflows compiled), and make fmt as instructed.

Verified locally on the merge commit: go build ./..., golangci-lint run ./pkg/workflow/... (0 issues), go test ./pkg/workflow/... (pass), and the relevant JS test suites (safe_outputs_mcp_schema_validation.test.cjs, safe_outputs_config.test.cjs) pass.

PR mergeable_state is now blocked (no longer conflicting/dirty) — this reflects pending/stale status checks on the new HEAD, which need a maintainer re-trigger since my push doesn't re-run CI. No unresolved review threads remain.