What problem does it solve? Code reviews often suffer from inconsistent severity labeling, missed security issues, and reviewers blocking merges over personal preference. This Skill turns code review into a repeatable workflow with a severity taxonomy, a 12-dimension checklist, and a structured verdict format. ## Core Features & Use Cases - Severity-tagged findings: Every comment is labeled Blocking, Should-fix, Nit, or Praise so authors instantly know what gates the merge. - Multi-dimension checklist: Walks the 12 Google eng-practices review dimensions plus concurrency and performance deep-dives via reference files. - OWASP security lens: Applies a security audit pass with concrete grep targets for injection, secrets, SSRF, deserialization, and auth issues. - Use Case: When asked to review a GitHub PR, the agent clones the change, reads every file in context, runs the checklist and security lens, then outputs a Summary, tagged Findings, and a verdict such as LGTM or Needs changes. ## Quick Start Ask the agent to review this pull request and produce a severity-tagged findings list with a merge verdict.