langchain-ai

code-review

Perform a structured code review of changes, checking for correctness, style, tests, and potential issues.

langchain-ai 28,558 3,999 Updated 5mo ago

Resources

1
GitHub

Install

npx skillscat add langchain-ai/deepagents/code-review

Install via the SkillsCat registry.

About this skill

This skill performs a structured review of code changes by evaluating correctness, style, test coverage, and security. It helps prevent bugs, side effects, and security vulnerabilities by validating work against a predefined checklist. An agent or developer should use this skill after implementing changes to ensure the code meets quality standards before delivery.

SKILL.md

Code Review Skill

Use this skill after implementing changes to validate your work before delivering.

Review Checklist

1. Correctness

  • Changes solve the original issue/task
  • No unintended side effects on existing functionality
  • Edge cases are handled
  • Error handling is appropriate (not excessive)

2. Code Quality

  • Code matches existing style and patterns
  • No unnecessary complexity or abstraction
  • Variable and function names are clear
  • No dead code, commented-out code, or TODOs left behind

3. Tests

  • New functionality has test coverage
  • Existing tests still pass
  • Tests cover both happy path and error cases
  • Tests are not brittle (don't test implementation details)

4. Safety

  • No hardcoded secrets or credentials
  • User input is validated at boundaries
  • No SQL injection, XSS, or command injection vectors
  • File operations use safe paths

Process

  1. Read each modified file end-to-end (not just the diff)
  2. Run the test suite: execute("python -m pytest -v")
  3. Run linters if available: execute("ruff check .")
  4. Run the bundled lint check: execute("python /skills/code-review/lint_check.py .")
  5. Check against each item in the review checklist
  6. If any issues found, fix them and re-review
  7. When everything passes, the review is complete

Helper Scripts

  • /skills/code-review/lint_check.py — Scans Python files for missing
    docstrings, long functions (>50 lines), and bare except: clauses. Run it
    via execute("python /skills/code-review/lint_check.py [path ...]").