Learning Taxonomy

About

Learning Taxonomy provides an expert framework for organizing and persisting session-specific knowledge within the Claude Code ecosystem. It standardizes the categorization of caveats, patterns, error fixes, and architectural decisions, ensuring that critical insights are stored in the correct hierarchical memory location—from project-level CLAUDE.md files to personal user rules. By automating the routing of technical debt, recurring bugs, and team conventions, it helps maintain a high-quality, evolving codebase context that improves Claude's performance and accuracy over time.

Key Features

• Path-specific rule generation using glob patterns and YAML frontmatter
• Standardized categorization for caveats, patterns, error fixes, and architectural decisions
• Automated priority calculation based on confidence levels and category types
• 2 GitHub stars
• Source analysis integration using git history to validate learning confidence
• Hierarchical routing across project, user, and enterprise memory scopes

Use Cases

• Persisting session 'gotchas' and workarounds to prevent recurring bugs
• Managing the balance between personal workflow preferences and team-shared project instructions
• Defining path-specific coding standards that Claude must follow for certain file types