What problem does it solve? Reviewing code changes consistently is hard: reviewers miss security holes, breaking contract changes, and failing validation gates, and review quality varies from person to person. This Skill standardizes code review by running the repository's configured validation commands, applying a built-in staff-engineer checklist plus any repo-local rules, and producing a structured, severity-ranked report with a clear approve or request-changes verdict. ## Core Features & Use Cases - Mandatory validation gate: Runs every command in the repo's validation.commands config in order and treats any failure as a blocker finding, so the review mirrors what CI will do. - Layered checklists: Applies a built-in checklist covering correctness, security, breaking changes, tests, concurrency, performance, and dependencies, extended by repo-local CODE_REVIEW.md, BACKWARD_COMPATIBILITY.md, and configured checklist files. - Structured verdict report: Produces a human-readable report with a validation-gate table, findings grouped by blocker/major/minor/nit severity with file and line references, a breaking-change checklist, and a mechanical approve/request-changes verdict. - Use Case: Point it at a pull request number before merging; it fetches the diff via tracker operations, runs the test and lint gates, flags a missing permission check as a blocker, and posts a full review report that maintainers can act on directly. ## Quick Start Ask the agent to review the current branch's changes against the base branch and produce a severity-ranked code review report with a merge verdict.
