Skip to content

CLAUDE.md

Latest commit

 

History

History
43 lines (30 loc) · 1.7 KB

CLAUDE.md

File metadata and controls

43 lines (30 loc) · 1.7 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

This is a reusable GitHub Actions workflow for AI-powered code review. It's designed to be called from other repositories via workflow_call, not run directly.

Architecture

The workflow (ai-code-review.yml) consists of 3 jobs:

  1. security-scan - Uses Trivy for multi-language vulnerability scanning (dependencies, code, IaC)
  2. ai-code-review - Uses AI (Gemini/OpenAI/Claude) to analyze PR diffs and score changes on 5 dimensions
  3. auto-approve-and-merge - Conditionally approves and merges PRs that pass AI review

Key design decisions:

  • Multi-AI provider support via ai-provider input (gemini, openai, claude)
  • Language-agnostic: no language-specific linting, Trivy handles all languages
  • AI review uses inline Python embedded in the workflow YAML
  • Supports Chinese (zh) and English (en) review output
  • Smart diff handling: auto-approves docs-only PRs, excludes sensitive files
  • Updates existing PR comments instead of creating duplicates
  • API calls include retry logic with exponential backoff

Input/Output Reference

Key Inputs:

  • ai-provider: gemini | openai | claude
  • ai-model: Custom model name (empty = provider default)
  • exclude-patterns: Comma-separated glob patterns to skip
  • custom-prompt: Additional review instructions

Secrets:

  • AI_API_KEY: API key for the selected provider

Testing Changes

This repository has no automated tests. To test workflow changes:

  1. Create a test repository that calls this workflow
  2. Open a PR in the test repository to trigger the workflow
  3. Check the workflow run logs in GitHub Actions