# Code Review Protocol > Structured code review across 5 dimensions — correctness, safety, patterns, coverage, scope. Verdict-routed to auto-fix or surface. Use when this capability is needed. - Skill: `tomevault-io/code-review-protocol` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/code-review-protocol` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/code-review-protocol/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/code-review-protocol --- # Code Review Protocol Five-dimension review. Input: diff, list of changed files. Output: verdict + routing. ## Dimensions 1. **Correctness** — Does it work? Logic errors, race conditions, off-by-one, type mismatches. 2. **Safety** — Edge cases, auth bypass, data leakage, injection, money-path errors. 3. **Patterns** — Follows project conventions? Naming, error handling, logging, module structure. 4. **Coverage** — Tests added/updated? Edge cases covered? Snapshot drift explained? 5. **Scope** — Touches only declared files? No drive-by changes. See `references/review-dimensions.md` for full rubric. ## Procedure 1. Load changed files and diff 2. Score each dimension: pass / needs_fix / needs_review 3. Map to action per `references/verdict-routing.md` 4. If `needs_fix` with suggested fix → dispatch fix-agent per routing table 5. If `needs_review` → write finding to inbox ## Verdict | Verdict | Meaning | |---------|---------| | pass | No issues found | | needs_fix | Specific fixable issue (with suggested change) | | needs_review | Structural concern, needs human judgment | --- > Source: [ansonphong/meta-dev](https://github.com/ansonphong/meta-dev) — distributed by [TomeVault](https://tomevault.io).