A context management system for AI-assisted development ✨
Track features with ledgers 📋 • Maintain coding patterns 🎯 • Keep AI assistants aligned with your project 🤖
Choose your preferred package manager:
# npm
npm install -g wrinkl
# pnpm
pnpm add -g wrinkl
# yarn
yarn global add wrinkl# Initialize in your project
cd my-project
wrinkl init
# Create a feature ledger
wrinkl feature user-authentication
# List active features
wrinkl list
# Archive completed features
wrinkl archive user-authentication🔥 ULTIMATE PROTIP: After running
wrinkl init, ask your AI coding assistant to automatically populate the entire.ai/directory for you! ✨
Wrinkl creates a .ai/ directory in your project with:
- 📄 Context files for AI assistants to understand your project
- 📚 Pattern documentation to maintain consistency
- 📋 Feature ledgers to track work progress
- 🏗️ Architecture decisions to guide development
- 🤖 Better AI Assistance - AI tools understand your project context
- 📝 Feature Tracking - Ledgers document progress and decisions
- 🎯 Pattern Consistency - Maintain coding standards across the team
- 🔄 Living Documentation - Context evolves with your project
Important: Keep your feedback loops tight! AI works better on focused tasks rather than sprawling features
"After 2+ years of coding exclusively with AI, I've learned that context is everything."
As a software engineer with 15 years of experience, I've witnessed the AI revolution transform how we build software. Wrinkl is my attempt to formalize the patterns and processes that make AI-assisted development truly effective.
The Problem: AI assistants are incredibly powerful, but they often lack the context needed to make decisions that align with your project's goals, patterns, and constraints.
The Solution: A structured approach to context management that keeps your AI assistants informed, your team aligned, and your codebase consistent.
This isn't just another tool—it's a methodology that evolves with the rapidly changing AI landscape.
💬 Want to chat about AI-assisted development? Hit me up on Discord: jayeeeffeff
After running wrinkl init, you'll have:
your-project/
├── .ai/
│ ├── README.md # Overview of the AI context system
│ ├── project.md # Project overview and requirements
│ ├── patterns.md # Coding patterns and conventions
│ ├── architecture.md # System architecture and decisions
│ ├── context-rules.md # Rules for AI assistants
│ └── ledgers/
│ ├── _active.md # Dashboard of active features
│ ├── _template.md # Template for new feature ledgers
│ ├── archived/ # Completed feature ledgers
│ └── [feature-name].md # Individual feature ledgers
├── .cursorrules # Cursor AI rules (optional)
├── augment.md # Augment AI context (optional)
└── .github/
└── copilot-instructions.md # GitHub Copilot instructions (optional)
Initialize the AI context system in your project.
Options:
-n, --name <name>- Project name (default: directory name)-t, --type <type>- Project type (default: "web app")-s, --stack <stack>- Technology stack (default: "TypeScript, Node.js")--no-cursor- Skip creating .cursorrules file--with-augment- Include augment.md file--with-copilot- Include GitHub Copilot instructions
Example:
wrinkl init --name "My App" --type "mobile app" --stack "React Native, Node.js"Create a new feature ledger to track development progress.
Example:
wrinkl feature user-authenticationList all active feature ledgers and their status.
Options:
-a, --all- Include archived features
Example:
wrinkl list --allArchive a completed feature ledger.
Example:
wrinkl archive user-authenticationThe .ai/project.md file contains your project's core information:
- Project goals and constraints
- Technology stack
- Key requirements
- Development workflow
The .ai/patterns.md file documents:
- Code style and conventions
- Common patterns and anti-patterns
- Testing strategies
- Performance guidelines
The .ai/architecture.md file captures:
- System design decisions
- Technology choices and trade-offs
- Scalability considerations
- Security architecture
Individual feature files track:
- Feature requirements and goals
- Technical approach and decisions
- Progress updates and blockers
- Testing and deployment notes
The .ai/context-rules.md file provides:
- Guidelines for AI assistants
- Code quality standards
- Security and performance rules
- Project-specific requirements
- Keep context updated 🔄 - Regularly update project files as requirements change
- Use feature ledgers 📝 - Create a ledger for each significant feature
- Document decisions 📋 - Record important technical decisions in ledgers
- Review patterns 🔍 - Regularly review and update coding patterns
- Reference context 📖 - Tell AI assistants to read the
.ai/directory - Mention features 🎯 - Reference specific feature ledgers when working
- Update progress ⏱️ - Keep ledgers updated with progress and decisions
- Follow patterns ✅ - Ensure AI-generated code follows project patterns
"I'm working on the user-authentication feature. Please read the feature
ledger in .ai/ledgers/user-authentication.md and help me implement the
login component following the patterns in .ai/patterns.md"
"Please review the project context in .ai/project.md and suggest an
architecture for the new notification system, documenting your decisions
in a new feature ledger"
If you use Cursor, the .cursorrules file provides context and guidelines for the AI assistant.
If you use Windsurf, simply rename .cursorrules to .windsurfrules - the content is identical, just different filename conventions.
The augment.md file provides context for Augment AI when working on your project.
The .github/copilot-instructions.md file guides GitHub Copilot to generate code that follows your project patterns.
- 🍴 Fork the repository
- 🌿 Create a feature branch
- ✏️ Make your changes
- 🧪 Add tests
- 📤 Submit a pull request
MIT - see LICENSE file for details.