The Core: The FASTER Framework
The project is built around Jim Kwik's FASTER methodology (Forget, Act, State, Teach, Enter, Review), implemented as an automated software layer. It moves beyond "watching tutorials" by forcing active engagement and retrieval.
Key Capabilities
- Claude Code Integration: Built as a specialized agent layer for Claude Code, adding
/learnand/reviewcapabilities directly to the developer's CLI. - Teach-Back Engine: Validates understanding by requiring the user to explain concepts back to the AI, which then identifies gaps in the mental model.
- Automated SRS: A built-in Spaced Repetition System that tracks your mastery levels and automatically schedules "Retrieval Expeditions" at optimal intervals.
- Adaptive Coaching Modes: Toggle between Balanced, Exam-Prep, Theory-Focused, and Practical modes to match your immediate learning objectives.
Technical Execution
- Stack: Python 3.12+
- Tooling: Built with uv for ultra-fast performance and zero-config execution.
- Agent Architecture: Leverages specialized Markdown-based prompting and structured
.learning/data persistence to maintain context across long-term learning journeys.
This kit is my response to the information deluge—a tool to ensure that in the age of AI, our human ability to master new skills remains our greatest competitive advantage.