What problem does it solve? When the Codex review bot is unavailable during the notion-dev review-and-merge flow, teams lose their automated code review gate. This Skill provides a fallback review rubric so a fresh, independent-context reviewer agent can still evaluate pull-request or branch diffs with consistent standards. ## Core Features & Use Cases - Severity-graded findings: Classifies issues as Critical, Required, Optional, Nit, or FYI, ordered by leverage so real defects are not buried under nits. - Machine-parseable verdict: Emits a strict output contract ending in VERDICT: CLEAN or VERDICT: NOT-CLEAN, enabling automated merge gating. - Selective review axes: Applies only relevant axes (correctness, readability, architecture, security, performance/robustness) based on the actual change, avoiding speculative or cosmetic findings. - Use Case: During a notion-dev feature flow, the Codex bot times out. A fresh reviewer agent applies this rubric to the PR diff, reports two Required findings with concrete fixes, and emits VERDICT: NOT-CLEAN so the orchestrator blocks the merge until fixed. ## Quick Start Ask the reviewer agent to review the current pull-request diff using the local code review rubric and report findings with a final verdict line.
