🤖 Bonjay — AI Skill Generator

Ever felt like your AI assistant (Cursor, Copilot, Gemini, Claude) just randomly blanks out or hallucinates because it doesn't understand your project's context? Or maybe you're tired of burning through tokens just to explain basic coding fundamentals over and over again?

Bonjay is here to fix exactly that! 🚀

Bonjay is a CLI tool that acts as the ultimate bridge between you and your AI agent. Think of it as a "pocket rulebook" for your AI. By scaffolding AI tooling context (rules & skills) directly into your project, your AI instantly understands your codebase and fundamental software engineering principles (like SOLID, GoF patterns, OWASP) without you having to type long, repetitive prompts.

Why use Bonjay?

• 🤝 Make Your AI Smarter: Significantly reduce AI hallucinations by providing clear, curated best practices and context from the get-go.
• 💸 Save Tokens & Time: Stop copy-pasting .cursor/rules/*.mdc rules or writing prompts from scratch. Let Bonjay scaffold the context in seconds!
• 🌍 Language-Agnostic Core: The fundamental rules shipped with Bonjay (Algorithms, Database Design, Security) apply universally.
• 🛡️ OWASP Security Standards: Integrated OWASP-based security principles for API design, including Mass Assignment prevention and Information Disclosure protection.
• 🎯 Multi-Agent Support: Output is perfectly optimized for Gemini, Cursor, and Claude with target-specific formatters.
• 🧠 Project Memory System: Automatically injects your project's history and decisions into the AI's context.
• 🚀 Rule Evolution: Capture new rules and lessons learned directly from your chat session.
• 🗂️ Single Source of Truth: Keep track of everything your AI has learned through the bonjay.json file.
• ⚡ Blazing Fast: Uses lazy loading and concurrent disk I/O to scaffold rules instantly, even on massive repositories.

Quick Start

Initialize Bonjay in your project:

npx bonjay init

This will launch an interactive wizard where you can select which AI platforms you use and what skill groups you want to inject into your project.

Available Skill Groups

Bonjay currently ships with 13 domains across 5 core groups:

1. 📦 CS Fundamentals (Big O, Data Structures, Algorithms)
2. 🏗️ Software Design (Design Patterns, Clean Architecture, Refactoring)
3. ⚙️ Engineering Practices (Code Quality/SOLID, Testing Strategy, Concurrency)
4. 🗃️ Data & Systems (Database Fundamentals, Query Optimization, API Design, Memory Management)
5. 🔒 Security (OWASP Top 10, AuthN/AuthZ, Secret Management, Mass Assignment Prevention)
6. 🧠 Project Context (Project Memory, Historical Context, Knowledge Items)
7. 🤖 AI Persona (Communication Style, Persona Rules)
8. 🔀 Version Control (Git Commits, Branching Strategy, Releases)

Commands

init

Interactive wizard to set up bonjay.json and generate your first set of skills.

bonjay init

add

Add a specific skill or an entire group to your project.

# Add a single skill
bonjay add software-design/design-patterns

# Add an entire group
bonjay add cs-fundamentals

remove

Remove a skill from your project and delete its generated files.

bonjay remove software-design/design-patterns

list

List all available and currently installed skills.

bonjay list

update

Regenerate all installed skills with the latest rules provided by your installed version of Bonjay.

bonjay update

capture

Capture a new rule or insight learned during development into your project's memory.

bonjay capture "Always use PascalCase for types" --title "Naming Convention"

How It Works

1. Bonjay maintains a library of curated Markdown rules with rich metadata.
2. When you run init or add, Bonjay writes to a bonjay.json configuration file in your project root.
3. It then generates the actual files (SKILL.md and rules/) in directories corresponding to your selected AI targets (e.g., .gemini/skills/).
4. Commit both bonjay.json and the generated rules to your repository.

Extensibility

Bonjay is built with a decoupled Registry system and supports Dynamic Skill Discovery. You can create custom skills specifically for your project by simply dropping your JavaScript modules (.js or .mjs) into the .bonjay/skills/ directory at the root of your project. Bonjay will automatically discover and load them at runtime without needing to recompile the core!

In the future, you will also be able to install custom skill packs from the community via npx bonjay add user/custom-skill.

Contributing

We welcome contributions from the community! Whether it's adding new skills, fixing bugs, or improving documentation, your help is appreciated.

Please read our Contributing Guidelines and our Code of Conduct before submitting a pull request.

License

This project is licensed under the MIT License - see the LICENSE file for details.