fix(skills): enforce executing-plans review gate (#41)

[ddddddddwp]

Fixes #41.

What changed

• Added an executing-plans review gate to Chinese and English comet-build skill docs.
• Required at least one code reviewer dispatch after all planned tasks complete and before the build -> verify guard runs.
• Required CRITICAL review findings to be fixed before verify, and accepted non-CRITICAL findings to be recorded with rationale and impact scope.
• Added prose regression assertions for the Chinese and English skill requirements.
• Fixed the Lingma init E2E test fixture so the mocked Superpowers staging install creates the expected .claude/skills payload before copying to Lingma.

Why

The comet-build + executing-plans path could previously finish implementation and transition to verify without any mandatory independent review. That made the lightweight execution path weaker than intended and allowed review to be skipped.

Validation

• npm run build
• npx vitest run test/ts/skills.test.ts test/ts/init-e2e.test.ts
• git diff --check origin/master...HEAD

---

Restores closed PR #70. The original PR could not be reopened because GitHub reports that the source repository for #70 had been deleted.

Summary by CodeRabbit

• Documentation

  • Added an executing-plans code-review gate: require a Superpowers review request before the build→verify guard; critical findings must be fixed, non‑critical findings may be accepted but require documented rationale and impact
  • Updated exit conditions to reflect review completion and resolution/recording requirements

• Tests

  • Extended unit tests for the new review requirements
  • Enhanced E2E tests for skills installation and mocked home directory handling

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: b5a31af7-0cbb-42c7-889c-03619230daa7

📥 Commits

Reviewing files that changed from the base of the PR and between 3b9efac and 185d708.

📒 Files selected for processing (1)

• test/ts/skills.test.ts

🚧 Files skipped from review as they are similar to previous changes (1)

• test/ts/skills.test.ts

---

📝 Walkthrough

Walkthrough

This PR documents a mandatory code review gate for comet-build's executing-plans mode and updates tests/mocks: load and invoke the Superpowers requesting-code-review skill before build→verify, require CRITICAL fixes, and record non-CRITICAL acceptance rationale in durable artifacts.

Changes

Executing-Plans Review Gate

Layer / File(s)	Summary
Bilingual skill documentation assets/skills-zh/comet-build/SKILL.md, assets/skills/comet-build/SKILL.md	Chinese and English comet-build SKILL.md add an executing-plans review gate: after planned tasks complete and before the build→verify guard --apply, load requesting-code-review and request review at least once; require CRITICAL findings fixed before verify; record non-CRITICAL acceptance rationale and impact scope in durable artifacts; exit conditions extended accordingly.
E2E test harness and regression assertions test/ts/init-e2e.test.ts, test/ts/skills.test.ts	E2E mock for npx skills --agent claude-code now stages a SKILL.md in the resolved cwd and returns an installed buffer; an init test runs under a mocked os.homedir(); Chinese and English workflow safeguard tests add assertions validating that when build_mode=executing-plans the flow invokes requesting-code-review at least once, includes a build → verify transition, and enforces CRITICAL findings fixed before non-CRITICAL handling.

🎯 3 (Moderate) | ⏱️ ~22 minutes

🐰 A rabbit hops through the fields of comet-build,
Where execution plans now heed the code review's call,
Critical findings fixed, non-critical reasons penned,
Each test stands guard, ensuring none slip through the gate,
A safer path to verify, bright and well-reviewed! 🌟

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: enforcing an executing-plans review gate in the comet-build skill, which is the primary focus of all file changes.
Linked Issues check	✅ Passed	All code requirements from issue #41 are met: executing-plans review gate added to build-phase documentation [#41], CRITICAL findings must be fixed and non-CRITICAL acceptances must be recorded [#41], verification tests validate the requirements [#41].
Out of Scope Changes check	✅ Passed	All changes are within scope: skill documentation updates for executing-plans requirements, E2E test fixture fixes for skill installation mocking, and regression tests validating the new requirements.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[ddddddddwp]

/gemini review

[coderabbitai]

Actionable comments posted: 3

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@docs/superpowers/plans/2026-06-03-executing-plans-review-gate.md`:
- Line 23: The heading "### Task 1: Add the Skill Safeguard Regression Test" is
an H3 under a top-level H1 with no H2 parent (MD001); fix by promoting the Task
header to an H2 (change "### Task 1: Add the Skill Safeguard Regression Test" to
"## Task 1: Add the Skill Safeguard Regression Test") or insert an appropriate
H2 parent section above it so the H3 has an H2 context; update the heading text
accordingly to preserve semantics and run the linter to confirm MD001 is
resolved.

In
`@openspec/changes/archive/2026-06-04-enforce-executing-plans-review/.astrolabe.yaml`:
- Line 15: Update the stale handoff_context value in .astrolabe.yaml so it
points to the archived location: replace the current
openspec/changes/enforce-executing-plans-review/.astrolabe/handoff/design-context.json
reference with the path under the archive
(openspec/changes/archive/2026-06-04-enforce-executing-plans-review/.astrolabe/handoff/design-context.json)
so consumers resolving the handoff_context key can find the real file; make this
change in the .astrolabe.yaml file where the handoff_context key is defined.

In `@openspec/changes/archive/2026-06-04-enforce-executing-plans-review/tasks.md`:
- Around line 2-3: Update the archived checklist so the two items are not
contradictory by making the first item describe a prior pending state and the
second item the completed follow-up; specifically, change the Line 2 text "Run
focused verification and record that English synchronization is pending user
confirmation." to a past-tense/chronological phrase such as "Ran focused
verification and recorded that English synchronization was pending user
confirmation at that step," and keep Line 3 "Sync the approved review gate to
`assets/skills/comet-build/SKILL.md` and cover it with English skill
assertions." as the completed action to make the final state unambiguous.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 5c0b41b7-5d37-4079-830c-dff0132d2b3a

📥 Commits

Reviewing files that changed from the base of the PR and between 5e8a6d7 and 791e243.

📒 Files selected for processing (16)

• assets/skills-zh/comet-build/SKILL.md
• assets/skills/comet-build/SKILL.md
• docs/superpowers/plans/2026-06-03-executing-plans-review-gate.md
• docs/superpowers/reports/2026-06-04-enforce-executing-plans-review-verify.md
• docs/superpowers/specs/2026-06-03-executing-plans-review-gate-design.md
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/.astrolabe.yaml
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/.astrolabe/handoff/design-context.json
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/.astrolabe/handoff/design-context.md
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/.openspec.yaml
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/design.md
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/proposal.md
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/specs/comet-build-review-gate/spec.md
• openspec/changes/archive/2026-06-04-enforce-executing-plans-review/tasks.md
• openspec/specs/comet-build-review-gate/spec.md
• test/ts/init-e2e.test.ts
• test/ts/skills.test.ts

[benym]

The openspec and superpowers directories should not be submitted. Please resolve this issue first, then we can proceed with the review.

[benym]

LGTM