Strike Logger 🎯

AI Code Review Strike Logger - Track and categorize AI-generated code errors

Transform unstructured code review feedback into structured learning data for AI coding assistants.

🚀 Quick Start

# Install globally
npm install -g @halonic/strike-logger

# Or use with npx
npx @halonic/strike-logger init

📖 Usage

Initialize in your project

cd your-project
strike-logger init

Log a strike from a commit

# Manual logging
strike-logger log --commit abc123 --message "Missing error handling in API call"

# Auto-detect from recent fix commits (analyzes the whole commit, all files)
strike-logger log --auto

# Interactive selection and review
strike-logger log --auto --interactive

List, resolve, and delete strikes

# List strikes with their ids (filter by category/severity/status)
strike-logger list
strike-logger list --open --severity high
strike-logger list --category null_pointer -n 50

# Mark a strike resolved (accepts a full id or an unambiguous id prefix)
strike-logger resolve 23df238a

# Delete a strike (prompts for confirmation, or pass --yes)
strike-logger delete 23df238a --yes

Auto-inject rules into active AI instruction files

# Auto-detect spec files (.cursorrules, .clinespec, etc.) and inject rules
strike-logger inject

Enable Git Pre-Commit Safeguards

# Register the hook so staged diff changes are scanned automatically before git commits
strike-logger hook --register

# To disable/unregister:
strike-logger hook --unregister

View statistics

# Terminal summary
strike-logger report

# Visual HTML dashboard (severity/category bars, monthly trend, recent strikes)
strike-logger report --html report.html

# Export raw data as CSV
strike-logger report --csv strikes.csv

CI/CD integration (GitHub Actions)

# Scaffold a workflow that fails PRs reintroducing known anti-patterns
strike-logger ci --init

# Run the check locally against a diff range (this is what CI runs)
strike-logger ci --check --base origin/main --head HEAD

Generate anti-pattern templates

# For ChatGPT/GPT-4
strike-logger generate --model gpt

# For Claude
strike-logger generate --model claude --output claude-antipatterns.md

# For GitHub Copilot
strike-logger generate --model copilot

🎯 Features

• 🔍 Auto-Detection: Automatically scans git history for fix commits (whole-commit, multi-file analysis)
• ✍️ Interactive Review: Review commits interactively to toggle strikes, verify diffs, and write custom comments
• 📊 Smart Categorization: 17 predefined error categories with rule-based matching and confidence scoring
• 🗂️ Strike Management: list, resolve, and delete strikes by id or unambiguous prefix
• 📈 Rich Reporting: Terminal summary, standalone HTML dashboard, and CSV export with monthly trends
• ⚙️ Custom Categories: Define your own rules, labels, and severity overrides via a config file
• 🤖 CI/CD Integration: Scaffold a GitHub Actions workflow that fails PRs reintroducing known anti-patterns
• 💉 Spec Injection: Auto-inject anti-pattern rules directly into .cursorrules, .clinespec, and other active AI configs
• 🛡️ Hook Safeguard: Git pre-commit hook scanning to block new commits containing known recurring bugs
• 💾 Local Storage: All data stored locally in JSON with atomic writes and schema migration (zero telemetry)
• 🔒 Privacy First: No external API calls, fully offline

📋 Commands

| Command | Description | |---------|-------------| | init | Initialize strike logger in current directory | | log | Log a new strike from commit or manual entry | | list | List logged strikes with their ids (filterable) | | resolve | Mark a strike resolved by id or unambiguous prefix | | delete | Delete a strike by id or unambiguous prefix | | report | View strike statistics; export --html / --csv | | generate | Generate anti-pattern prompt templates | | inject | Auto-inject generated anti-pattern rules into active spec files | | hook | Manage git pre-commit hook safeguard | | ci | Scaffold a CI workflow (--init) or check a diff range (--check) |

⚙️ Custom Configuration

Drop a strike-logger.config.json (or .strikerc.json) in your project root to extend categorization without touching source. A bad config is ignored, never fatal.

{
  "customRules": [
    {
      "category": "flaky_test",
      "keywords": ["flaky", "sleep"],
      "patterns": ["setTimeout.*test", "\\bflaky\\b"],
      "weight": 1.2
    }
  ],
  "categoryLabels": { "flaky_test": "Flaky Tests" },
  "severityOverrides": { "flaky_test": "high" }
}

• customRules — extra rules merged after the built-ins (patterns are case-insensitive regex strings).
• categoryLabels — human-readable names shown in reports.
• severityOverrides — force a severity for a category, taking precedence over the heuristics.

🗂️ Strike Categories

The tool automatically categorizes errors into:

Code Quality

• Missing Error Handling
• Null Pointer Exceptions
• Race Conditions
• Memory Leaks

Security

• SQL Injection
• XSS Vulnerabilities
• Hardcoded Secrets
• Auth Bypass

Architecture

• Tight Coupling
• Missing Abstractions
• Wrong Design Patterns

Performance

• N+1 Query Problems
• Inefficient Algorithms

Testing

• Missing Tests
• Insufficient Coverage

And more...

💡 Example Workflow

# 1. Initialize
cd my-ai-project
strike-logger init

# 2. Auto-scan recent fixes
strike-logger log --auto

# 3. View what went wrong
strike-logger report

# 4. Generate anti-patterns for your AI
strike-logger generate --model gpt > ai-guidelines.md

# 5. Use guidelines in your AI prompts!

🎨 Sample Output

📊 STRIKE REPORT

Total Strikes: 23

🎯 By Severity:
  Critical: 2
  High: 8
  Medium: 10
  Low: 3

📁 By Category:
  missing_error_handling: 7
  null_pointer: 5
  missing_tests: 4
  ...

🛠️ Development

# Install dependencies
npm install

# Build
npm run build

# Test
npm test

# Development mode
npm run dev

📦 Project Structure

strike-logger/
├── src/
│   ├── cli.ts              # CLI entry point
│   ├── parsers/            # Git diff parsing (multi-file, diff ranges)
│   ├── categorizer/        # Error categorization + confidence scoring
│   ├── config/             # User config loader & custom-rule compilation
│   ├── database/           # Strike storage, migration, atomic writes
│   ├── templates/          # Template generation and auto-injector
│   ├── reporting/          # HTML dashboard generation
│   ├── ci/                 # GitHub Actions workflow + diff-range check
│   ├── interactive/        # Interactive CLI prompting
│   └── hooks/              # Git hook + shared anti-pattern scanner
├── data/
│   └── strikes.json        # Local strike database
└── tests/                  # Vitest unit test suites

🚀 Roadmap

• [ ] VS Code Extension (real-time strike detection)
• [ ] Team Dashboard (centralized analytics)
• [x] Git Hook Pre-Commit Integration
• [x] Auto-Spec / Prompts rules injection
• [x] CI/CD Integration (GitHub Actions)
• [x] Custom category definitions
• [x] HTML/CSV reporting with trends
• [x] Strike management (list / resolve / delete)
• [ ] ML-based categorization

📄 License

MIT © Halonic

🤝 Contributing

Contributions welcome! This is an open-source project from Halonic.

Built with ❤️ by Halonic | Report Issues