pr-review
Fix engine for PR review comments. Fetches review comments (Gemini bot or human), categorizes by impact, posts a prioritized fix queue, and applies fixes on dev approval. Called directly for quick fixes, or internally by pr-review-agent as part of full PR review.
git clone --depth 1 https://github.com/wednesday-solutions/ai-agent-skills /tmp/pr-review && cp -r /tmp/pr-review/skills/pr-review ~/.claude/skills/pr-reviewSKILL.md
# PR Review — Gemini Fix Queue ## Trigger Load this skill when a dev wants to **act on PR review comments**: - "Fix the review comments" - "Triage the PR feedback" - "Apply fixes from the review" - "@agent fix #1 #3" - "@agent fix all" - "What do I need to fix in this PR?" Also triggered automatically by GitHub Actions when Gemini bot posts a PR review. **Do NOT use this skill for:** creating a new PR (use `pr-create`), or committing code (use `git-os`). This skill only runs on an already-open PR that has review comments on it. ## Priority Order Fix in this order — lower number = fix first. | Rank | Category | Examples | |------|----------|---------| | 1 | security | auth issues, injection risks, data exposure | | 2 | breaking | API contract changes, interface changes | | 3 | logic | wrong conditions, missing edge cases | | 4 | performance | N+1 queries, unnecessary re-renders | | 5 | naming | variable/function/class names, casing | | 6 | style | formatting, whitespace, import order | **Rule: never fix a style item while a security or breaking issue is pending.** ## Review Report Format ```markdown # Gemini Review — PR #<n> | # | Category | File | Issue | Status | |---|----------|------|-------|--------| | 1 | security | src/db.js | SQL query not parameterized | ⬜ pending | | 2 | logic | src/user.js | Missing null check on user.profile | ⬜ pending | | 3 | naming | src/auth.js | Variable `x` is unclear | ⬜ pending | To fix: `@agent fix #1 #2` Fix all: `@agent fix all` ``` ## Tools | Action | Tool | |--------|------| | Read a file before applying a fix | `Read` | | Apply a fix to a file | `Edit` | | Run git commands (commit, push) | `Bash` | | Search for a pattern across files | `Grep` | | Find files by name | `Glob` | ## Agent Fix Rules - Never auto-fix without explicit dev approval (`@agent fix #N`) - Read `git-os` SKILL.md before making any commit - One commit per fix item - Commit format: `fix(scope): description\n\nResolves review item #N` - Push to the same PR branch - Update the report — mark fixed items as `✅ fixed` ## Failure Handling If a fix cannot be applied cleanly, post a comment explaining the conflict. Never force-push or silently skip a fix.
Use when asked to audit a module, check its health, or assess whether it should be refactored. Runs structural query, risk check, and test generation automatically.
Use when a dev is new to the project, asks for an overview, or wants to get oriented quickly. Fills graph gaps then runs a guided codebase interview.
Full PR review orchestrator. Use when asked to review a PR or check if it is ready to merge. Runs blast radius + drift check on changed files, then hands off to pr-review for comment triage and fixes. Do NOT use pr-review directly for full reviews — use this.
Enforces architecture boundaries defined in PLAN.md. Use when a PR crosses module/service boundaries, when the dev asks "are we following the architecture?", or as a scheduled architecture health check. Not for querying what a module does — use brownfield-chat for that.
Unified codebase intelligence. Handles all questions about structure, logic, risk, and dependencies. Combines natural-language Q&A with deterministic lookups and pre-edit blast radius checks.
Pre-deploy and post-deploy checklist skill. Ensures env vars, migrations, CI, rollback plan, smoke tests, and monitoring are verified before and after every deployment.
Parallel persona planning for new projects. Research agent runs first to build domain context, then Architect, PM, and Security agents run in parallel. Synthesis agent combines all perspectives into a detailed GSD-style PLAN.md with Tensions section.
Unified development and design standards. Enforces code quality (complexity < 8), strict naming conventions, and the mandatory use of approved UI component libraries.