Skip to content
Back to skills

Code Review Skill

ASecurity

Provide structured software code reviews that confirm scope, focus on root causes, evaluate tests, and enforce project conventions.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 27, 2026
code-qualitycode-reviewgitperformancedocumentation

Security analysis

A100/100

Scanned September 27, 2026

npx -y skills add David-Li0406/meta-skill-evloving --skill code-review-skill --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Code Review Skill?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Code Review Skill
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/david-li0406-code-review-skill/badge)](https://www.skillsdirectory.com/skills/david-li0406-code-review-skill)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: code-review-skill
description: Provide structured software code reviews that confirm scope, focus on root causes, evaluate tests, and enforce project conventions.
---

# Code Review Skill

Use this when the user requests a review of code, pull requests, or diffs. Load the smallest relevant context first (changed files, linked tickets, reproduction steps) before drilling into implementation details.

## Review flow

1. **Establish scope**
   - Summarize the change intent, affected components, and linked issues.
   - Ask for missing diffs, runtime configuration, or reproduction instructions.
2. **Trace the cause**
   - Important: try to fix things at the cause, not the symptom. Follow data flow upstream to ensure the change touches the real source of the problem.
   - Flag quick fixes that mask deeper defects or debt.
3. **Manage context**
   - Pin key files and collapse unrelated noise.
   - Record assumptions and request clarification when signals conflict.
4. **Validate correctness**
   - Inspect logic paths, edge cases, and concurrency or performance concerns.
   - Map existing tests to behaviors; specify new tests or instrumentation when coverage is thin.
5. **Check conventions**
   - Compare against architecture and style guides. Surface deviations with concrete alternatives or references.
6. **Communicate outcomes**
   - Grade findings by severity (blocker, follow-up, nit).
   - Offer actionable edits, test commands, or documentation updates.

## Response template

```
### Findings
- [severity] file:line — issue + why it matters

### Questions
- Clarifications needed to proceed

### Recommended actions
- Ordered list of fixes, tests, or follow-ups
```

## Safeguards

- If critical context is missing, pause and ask for it instead of guessing.
- Never suggest destructive git commands unless the user has already used them in the conversation.
- When advising on tests, prefer deterministic automated coverage; call out manual validation only when unavoidable.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…