claude-code-memory

What problem does it solve?

As AI interactions grow, Claude's internal memory (.claude/memory/) can become cluttered with stale, redundant, or conflicting information, leading to context pollution and reduced effectiveness. This Skill automates the process of auditing, organizing, and updating Claude's memory, ensuring it remains concise, accurate, and highly relevant for every session. This saves time by preventing Claude from repeating past struggles and improves the quality of its responses.

Core Features & Use Cases

• Automated Memory Audit: Quickly scan memory files for staleness, excessive size, redundancy, and organizational issues, providing a clear report of areas needing attention.
• Structured Learning Capture: Systematically document solutions to problems Claude struggled with, preventing the AI from repeating the same mistakes in future sessions.
• Memory Lifecycle Management: Provides workflows to promote temporary "learnings" into permanent "conventions," "architecture," or "decisions," ensuring knowledge evolves with the project.
• Use Case: After a complex debugging session where Claude struggled with a specific Python import error, the user can activate this skill to capture the problem and its solution. In future sessions, Claude can reference this learning to quickly resolve similar issues, saving significant time and effort.

Quick Start

To get an initial health check of Claude's memory files: python scripts/audit_memory.py To capture a new learning after solving a difficult problem: python scripts/capture_learning.py --quick --category debug --title "Python import error fix" --problem "Module not found despite correct structure" --solution "Used python -m module.main"