pr-review
Reviews a GitHub pull request and posts inline comments plus one consolidated summary, adapting to any codebase by discovering the project's own test runner, requirement specs, and architecture conventions before running six specialized review agents in parallel. Stack-agnostic across language and framework; targets GitHub PRs via the gh CLI. Use when the user says "review PR 128", "review this PR", "code review this PR", or "check this pull request". Do NOT use for creating PRs or responding to review comments (use gh-address-comments), or debugging failing CI checks (use gh-fix-ci).
npx skills add tech-leads-club/agent-skills --skill pr-review --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Coordinates 6 specialized subagents to perform a PR review in parallel, without analyzing the diff itself. It discovers project-specific rules (test runner, requirements, architecture conventions, and local review skills) and applies those rules via the subagents. It then consolidates findings into one PR-level summary and posts inline comments only through the gh CLI, avoiding direct inline analysis within the orchestrator. It is designed to work on stack-agnostic codebases and requires explicit PR context (repo, PR number, SHA) and a discovery map derived from the target repository.
How it works
- Acts as the orchestration layer (Step 1: Initialize) to gather PR context, repository, head SHA, existing inline comments, and PR intent.
- Performs project discovery to build a DISCOVERY MAP from the repository that describes test runners, requirements sources, architecture conventions, and project-local review skills.
- Step 2 launches six subagents in parallel, each receiving the full context: {REPO}, {PR}, {SHA}, the diff, existing comments, PR intent, and the DISCOVERY MAP. Each subagent is self-contained and does not share context with the orchestrator.
- Step 3 would consolidate findings and post a PR-level summary, plus inline comments following strict posting rules (e.g., using temp body files for comment bodies, correct gh CLI endpoints).
- Enforces a strict protocol: do not analyze the diff or generate findings directly in the orchestrator; do not inline-review; post only via specified gh CLI commands; maintain per-line and per-comment uniqueness to avoid duplicates; and require a minimum confidence before reporting findings.
When to use it
Use when a user asks to review a PR (e.g., "review PR 128", "review this PR", "code review this PR", or "check this pull request"). Do not use for creating PRs, responding to review comments, or debugging CI checks (those are handled by other skill pathways).
What it can touch
- The PR context and metadata loaded from the repository.
- The DIFF content of the PR, the head SHA, and existing inline comments to avoid duplicates.
- The project discovery map which is passed verbatim to all subagents.
- It relies on the six subagents and the consolidation process to generate and post comments via gh CLI commands.
Caveats
- The skill is orchestration-only: it must not analyze the diff or generate findings itself.
- It uses the gh CLI for all fetches and postings.
- It requires the environment to support parallel task execution to launch six subagents in a single batch.
- It never auto-triggers during coding; explicit invocation is required.
# PR Review — Adaptive Orchestration Protocol Coordinates 6 specialized subagents (via the Task tool), then consolidates their findings into one summary. This skill is **stack-agnostic**: it owns no rules of its own. Instead it discovers what the project already documents — test runner, requirement specs, architecture conventions, project-local review skills — and each subagent loads and applies **whatever that project actually has**. It never invents stack-specific rules and never duplicates project docs. ## Execution Contract (NON-NEGOTIABLE) This skill is an **orchestration-only** protocol. Read these rules before anything else: 1. You are the **orchestrator**. You MUST NOT analyze the diff, read source files, or write any review finding yourself. Your only jobs are gathering context (Step 1), launching subagents (Step 2), and spawning the consolidation subagent (Step 3). 2. The six reviews MUST run as **six separate, general-purpose subagents** (no restricted toolset), launched **in parallel in a single batch**. Use whatever subagent / parallel-task mechanism your harness provides (in Claude Code, the Task tool). 3. Doing the review inline — even partially, even "to save time",
- Execution Contract (NON-NEGOTIABLE)
- Step 1: Initialize
- 1a. PR context
- 1b. Project discovery (the adaptive spine)
- Step 2: Launch subagents in parallel
- Severity labels (all subagents use these)
- Universal Rules (every subagent + consolidation must follow)
- Comment posting commands (copy exactly)
- Subagent 1: Security — <!-- pr-review:security -->
- Subagent 2: Requirements & Definition of Done — <!-- pr-review:requirements -->
- Track A — Issue tracker
- Track B — In-repo spec files
- Resolution logic
- Subagent 3: Test Coverage — <!-- pr-review:tests -->
gh api repos/{REPO}/pulls/{PR}/comments \
gh api repos/{REPO}/pulls/{PR}/comments/{COMMENT_ID}/replies -F body=@body.md
gh pr review {PR} --comment --body-file body.md
gh api repos/acme/app/pulls/12/comments \
gh api repos/acme/app/pulls/12/comments -f body=@body.md -f path=src/x ...What does the pr-review skill do?
Reviews a GitHub pull request and posts inline comments plus one consolidated summary, adapting to any codebase by discovering the project's own test runner, requirement specs, and architecture conventions before running six specialized review agents in parallel. Stack-agnostic across language and framework; targets GitHub PRs via the gh CLI. Use when the user says "review PR 128", "review this PR", "code review this PR", or "check this pull request". Do NOT use for creating PRs or responding to review comments (use gh-address-comments), or debugging failing CI checks (use gh-fix-ci).
How do I install it?
Run `npx skills add tech-leads-club/agent-skills --skill pr-review --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From tech-leads-club/agent-skills, a repository with 4,983 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.