Automates the extraction of coding patterns and lessons learned to create a self-improving project knowledge base.
This skill transforms Claude Code into a self-evolving agent by systematically capturing insights from every development session. It implements a structured framework to categorize successful approaches, corrections, and anti-patterns using dynamic confidence scoring. By formalizing knowledge acquisition into a machine-readable format and integrating it with project memory files like CLAUDE.md, it ensures that previous mistakes are avoided and high-performance coding habits are consistently reinforced across the codebase.
Key Features
01Automated pattern extraction for successes, corrections, and anti-patterns
02Structured YAML-based learning entries for reusable knowledge
03Standardized session wrap-up protocol for consistent memory maintenance
