# Code Review > Systematic code review with severity-ranked findings Use when this capability is needed. - Skill: `tomevault-io/code-review-210` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/code-review-210` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/code-review-210/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - 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-210 --- ## When to Use Use this skill when reviewing code changes (diffs, PRs, or complete files) for correctness, maintainability, security, and adherence to best practices. ## Steps 1. **Understand context**: Read the full diff or file. Identify the purpose of the change (feature, bugfix, refactor, etc.). 2. **Check correctness**: Verify logic, edge cases, error handling, and return values. Look for off-by-one errors, null/undefined handling, and race conditions. 3. **Assess security**: Check for injection vulnerabilities (SQL, XSS, command), improper input validation, hardcoded secrets, and insecure defaults. 4. **Evaluate design**: Consider naming clarity, function cohesion, abstraction level, and adherence to existing patterns in the codebase. 5. **Review tests**: Check if changes are covered by tests. Identify missing test cases for edge cases and error paths. 6. **Rank findings**: Assign severity to each finding and present them in priority order. ## Severity Levels - **Critical**: Bugs that will cause crashes, data loss, or security vulnerabilities in production. - **Warning**: Issues that may cause problems under certain conditions or violate important conventions. - **Suggestion**: Improvements to readability, naming, or style that are not strictly required. - **Nitpick**: Minor style preferences with no functional impact. ## Rules - Focus on the changed code, not unrelated pre-existing issues (unless they interact with the change). - Provide concrete fix suggestions, not just problem descriptions. - Be specific: reference line numbers, function names, and variable names. - Acknowledge good patterns and well-written code, not just problems. - If you're uncertain whether something is a bug, say so and explain your reasoning. ## Output Format For each finding: ``` [SEVERITY] file:line — Short title Description of the issue and why it matters. Suggested fix or alternative approach. ``` End with a summary: total findings by severity, overall assessment, and whether the change is ready to merge. --- > Source: [definableai/definable.ai](https://github.com/definableai/definable.ai) — distributed by [TomeVault](https://tomevault.io).