What problem does it solve? Reviewing pull requests thoroughly is time-consuming and inconsistent: small diffs get over-reviewed, large diffs get skimmed, and security issues slip through. This Skill automates context-aware PR review on GitHub, detecting whether you are reviewing your own draft PR or someone else's code and running the right workflow for each. ## Core Features & Use Cases - Dual review modes: Self-review mode runs review agents, posts findings, applies approved fixes, pushes commits, and optionally marks the PR ready; review mode posts inline comments and submits an approve/comment/request-changes decision without editing code. - Diff-scaled agent fan-out: Measures changed lines and files to pick a review tier, from inline review for tiny diffs up to six parallel agents (code-reviewer, silent-failure-hunter, pr-test-analyzer, comment-analyzer, type-design-analyzer, plus a security pass) for large ones. - Security-first review: Every tier includes a security-focused pass checking for injection, XSS, auth flaws, secrets, SSRF, and other exploitable vulnerabilities, with content-based overrides for auth, crypto, migrations, and shared workflows. - Use Case: After pushing a draft PR, run the skill with no arguments to self-review your branch, apply critical and important fixes automatically, and mark the PR ready for human review. ## Quick Start Ask the AI to review the pull request for the current branch, or pass a PR number such as #42 to review a specific pull request.
