name: commit-analysis description: | Automated commit analysis and regression detection for AItrader. 自动化提交分析和回归检测。

Use this skill when:

• Running regression checks after code changes (代码修改后运行回归检测)
• Analyzing git commit history (分析 git 提交历史)
• Using AI to analyze code changes (使用 AI 分析代码变更)
• Checking if fix commits are still correctly applied (检查修复是否正确应用)
• Before merging PRs (合并 PR 前)

Keywords: commit, regression, analysis, git, fix, validation, AI, 提交, 回归, 分析

Commit Analysis & Regression Detection

Overview

Automated tools for detecting regressions and analyzing code changes:

Recommended Tool: smart_commit_analyzer.py

This is the primary regression detection tool. Rules are automatically generated from git history.

Commands

# Full analysis (update rules + validate)
python3 scripts/smart_commit_analyzer.py

# Update rules only (scan git for new fixes)
python3 scripts/smart_commit_analyzer.py --update

# Validate only (check existing rules)
python3 scripts/smart_commit_analyzer.py --validate

# Show all rules
python3 scripts/smart_commit_analyzer.py --show-rules

# JSON output (for CI/CD)
python3 scripts/smart_commit_analyzer.py --json

Expected Output

🔍 Smart Commit Analyzer
============================================================

Step 1: 从 Git 历史更新规则库...
📊 扫描到 78 个修复提交
✅ 新增 5 条规则

Step 2: 验证所有规则...

============================================================
📋 验证结果
============================================================
✅ 通过: 70
❌ 失败: 2
⚠️  警告: 3
⏭️  跳过: 0

How It Works

┌─────────────────────────────────────────────────────────┐
│  Step 1: git log --grep="fix"                           │
│          → Auto-discover all fix commits                │
└─────────────────────────────────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────┐
│  Step 2: git show <commit> -- <file>                    │
│          → Extract key code patterns from diffs         │
└─────────────────────────────────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────┐
│  Step 3: Save to configs/auto_generated_rules.json      │
│          → Rules auto-grow with new fix commits         │
└─────────────────────────────────────────────────────────┘
                            ↓
┌─────────────────────────────────────────────────────────┐
│  Step 4: Validate all rules                             │
│          → Detect regressions (missing patterns)        │
└─────────────────────────────────────────────────────────┘

AI Deep Analysis (Optional)

Requires DEEPSEEK_API_KEY environment variable.

# Analyze last 10 commits with AI
python3 scripts/analyze_commits_ai.py --commits 10

# JSON output
python3 scripts/analyze_commits_ai.py --commits 10 --json

Git History Analysis

# Analyze last 50 commits
python3 scripts/analyze_git_changes.py

# Show only fix commits
python3 scripts/analyze_git_changes.py --fix-only

# Analyze more commits
python3 scripts/analyze_git_changes.py --commits 100

GitHub Actions Integration

These tools run automatically on every push/PR via .github/workflows/commit-analysis.yml:

Key Files

When to Run

• Before committing: python3 scripts/smart_commit_analyzer.py
• Before merging PR: Automatic via GitHub Actions
• After pulling updates: python3 scripts/smart_commit_analyzer.py --validate
• Investigating regressions: python3 scripts/smart_commit_analyzer.py --show-rules

Interpreting Results