Code Review Skill

SkillDev tools

Code review for Python changes in NexRL, following project conventions.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Code Review Skill skill

What this skill tells your AI

The instructions your AI receives, as published by nex-agi/nexrl in .claude/skills/code-review/SKILL.md and read by ahel’s review.

Overview

Review Python code changes following NexRL project conventions and .ai-instructions/code-review/generic.md.

Checklist

Code Quality

  • Follows project style (black, isort formatting)
  • No debug code or commented-out sections
  • No TODOs/FIXMEs unless documented
  • License headers present on Python/shell/YAML/TOML files
  • pylint and mypy pass without new warnings

Architecture & Design

  • Consistent with NexRL's loosely-coupled architecture
  • Proper use of existing abstractions (workers, trainers, services)
  • No circular imports
  • Configuration changes follow existing patterns (Hydra/OmegaConf)

Documentation Alignment

  • Documentation in docs/ reflects code changes
  • Docstrings present for public APIs
  • Configuration reference updated if parameters changed
  • Examples still work

Testing

  • Tests added for new functionality (in tests/)
  • No ad-hoc test files outside tests/
  • Existing tests still pass

Commit Content

  • Only relevant changes included
  • No build artifacts or sensitive information
  • Commit message follows conventions (present tense, no AI co-author)

How to Run

# Review staged changes
git diff --staged

# Review against base branch
git diff origin/main...HEAD

# Check specific files
python -m pylint <files> --rcfile=.pylintrc -rn -sn
python -m mypy <files> --ignore-missing-imports

Signals

GitHub stars
119
Forks
9
Last commit
Jul 2026
Advanced
Catalog kind
skill
Gateway key
code-review-nex-agi
Source
github.com/nex-agi/nexrl