# Pre Review

> Pre Review is an Agent Skill for testing & quality published in the arklexai/Agent-First-Organization repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Check your branch before opening a PR (lint, format, tests, title)

- Skill name: `pre-review`
- Category: [Testing & Quality](https://www.openparable.com/skill/testing-quality/1)
- Source: https://github.com/arklexai/Agent-First-Organization/blob/HEAD/.claude/skills/pre-review/SKILL.md
- Repository: arklexai/Agent-First-Organization (700 GitHub stars when collected)

## SKILL.md

# Pre-review

Run local checks to catch issues before opening a PR.

## Steps

### 1. Check branch and changes

```bash
git branch --show-current
git status
git log --oneline main..HEAD
```

### 2. Run ruff linter

```bash
ruff check .
```

If there are fixable issues, run `ruff check --fix .` and report what was fixed.

### 3. Run ruff formatter

```bash
ruff format --check .
```

If files would be reformatted, run `ruff format .` and report what changed.

### 4. Run tests

```bash
pytest tests/ -x --tb=short
```

Report pass/fail count and any failures.

### 5. Check coverage

```bash
pytest tests/ --cov=arklex --cov-fail-under=45 --cov-report=term-missing -q
```

Coverage minimum is 45%. Flag if below threshold.

### 6. Validate branch name

Branch should follow `<type>/<short-description>` pattern, e.g. `feat/add-retry`.

### 7. Suggest PR title

Based on the commits on this branch, suggest a Conventional Commits title:
- Format: `<type>(<scope>): <description>`
- Max 72 characters
- Types: feat, fix, docs, chore, ci, build, refactor, test, perf, style, revert

### 8. Report

Summarize results:

```
Lint:     pass/fail
Format:   pass/fail
Tests:    X passed, Y failed
Coverage: XX% (minimum 45%)
Branch:   <name> (valid/invalid)
Title:    <suggested title>
```

---

Source: https://www.openparable.com/skill/testing-quality/pre-review (OpenParable). Cite the canonical URL above.
