Align the seven plan-reviewer-*.md agent definitions in the plan-agent plugin with the proven product-plans sibling pattern: change the frontmatter key from allowed-tools: to tools: , narrow Bash to Bash(git *) , and append the "Teammate-only" clause to every agent description. Bump plan-agent to 1.10.0 as a PATCH fix for functional correctness.
Read and implement all steps in the plan at docs/plans/fix-review-plan-agent-frontmatter.md — Fix review-plan agent frontmatter and descriptions to match product-plans pattern. Verify against the plan's Tests, Verification, and Acceptance Criteria before reporting done. If everything passed, mark completion in docs/plans/fix-review-plan-agent-frontmatter.md — tick each step's [x] marker and each criterion's - [x], set status: completed — and re-render the HTML from the spec. If any check failed, leave status: in-progress and say which.
More ways to run this plan — goal & workflow prompts, file path
Achieve this goal: Fix review-plan agent frontmatter and descriptions to match product-plans pattern. The plan at docs/plans/fix-review-plan-agent-frontmatter.md describes one approach — use it as reference, but optimize for the outcome. Fan out across parallel subagents where that serves the outcome. Verify against the plan's Tests, Verification, and Acceptance Criteria before reporting done. If everything passed, mark completion in docs/plans/fix-review-plan-agent-frontmatter.md — tick each step's [x] marker and each criterion's - [x], set status: completed — and re-render the HTML from the spec. If any check failed, leave status: in-progress and say which.
Run a workflow to implement the plan at docs/plans/fix-review-plan-agent-frontmatter.md — Fix review-plan agent frontmatter and descriptions to match product-plans pattern. Brief subagents with the plan file at docs/plans/fix-review-plan-agent-frontmatter.md. Reserve a final verification phase for the lead agent, not a subagent. Verify against the plan's Tests, Verification, and Acceptance Criteria before reporting done. If everything passed, mark completion in docs/plans/fix-review-plan-agent-frontmatter.md — tick each step's [x] marker and each criterion's - [x], set status: completed — and re-render the HTML from the spec. If any check failed, leave status: in-progress and say which.
fix-review-plan-agent-frontmatter.html
docs/plans/fix-review-plan-agent-frontmatter.html
docs/plans/fix-review-plan-agent-frontmatter.md
Context
The story behind this plan — what prompted the work and why it matters now.
The review-plan Agent Team skill (v1.9.0) shipped with seven plan-reviewer-*.md agent definitions under kit/plugins/plan-agent/agents/ . A post-ship team review ( add-plan-review-team-skill-review.html ) identified two functional correctness issues and one convention gap by comparing the shipped files against the established product-plans sibling agents:
Wrong frontmatter key: All seven agents use allowed-tools: in their YAML frontmatter. The agent runtime reads tools: — allowed-tools: is the correct key for skills ( SKILL.md ), not agents. Every product-plans agent uses tools: . The restriction is likely silently ignored as-is.
Unrestricted Bash: The agents grant bare Bash access, which gives reviewers latent write capability. The product-plans agents use Bash(git *) to restrict Bash to read-only git commands.
Missing "Teammate-only" clause: The product-plans agents all end their descriptions with "Teammate-only — designed to run inside an Agent Team led by the plan-review-agents skill; not for standalone invocation." The plan-agent reviewer agents omit this, meaning they may auto-activate on broader intent matches than intended.
This is a PATCH fix (correctness of existing metadata, no new behavior) per marketplace.md versioning rules. The version bump is 1.9.0 → 1.10.0 since the frontmatter key change affects runtime behavior.
Files that change
Every file this plan touches, and what happens to each one.
.claude-plugin/marketplace.jsonmodified bump plan-agent 1.9.0 → 1.10.0kit/plugins/plan-agent/CHANGELOG.mdmodified add 1.10.0 entry- agents/
plan-reviewer-architecture.mdmodified frontmatter key + Bash restriction + teammate clauseplan-reviewer-completeness.mdmodified frontmatter key + Bash restriction + teammate clauseplan-reviewer-testability.mdmodified frontmatter key + Bash restriction + teammate clauseplan-reviewer-risk.mdmodified frontmatter key + Bash restriction + teammate clauseplan-reviewer-conventions.mdmodified frontmatter key + Bash restriction + teammate clauseplan-reviewer-ux.mdmodified frontmatter key + Bash restriction + teammate clauseplan-reviewer-accessibility.mdmodified frontmatter key + Bash restriction + teammate clause
Steps
The step-by-step work, in order — each step says what to do, why it matters, and how to check it worked.
Tests
The tests that prove the change does what it promises.
Definition of done
The plan counts as done when every statement below is true — check each one off as you verify it.
Final check
One last pass to confirm the whole change works end to end.
Structural parity check: Run head -6 kit/plugins/plan-agent/agents/plan-reviewer-*.md and head -6 kit/plugins/product-plans/agents/product-reviewer-*.md side by side. The frontmatter shape should now match: both use tools: , both grant Bash(git *) , both have a model: field, and both have descriptions ending with a "Teammate-only" clause.
No regressions: The agent body content (mandate, how-to-review, report-back sections) must be byte-identical to the v1.9.0 versions — only the frontmatter block changes.
JSON validity: node -e "JSON.parse(require('fs').readFileSync('.claude-plugin/marketplace.json'))" exits 0.
Wrapping up
Three gates that must all pass before this plan is marked completed.
Completion Report
No items to report — all requirements met.