Where can I see the results of the audit?

Detailed results, including a breakdown of all files and required remediation steps, are automatically written to .claude/audit/memory.md.

What is a CLAUDE.md memory file?

CLAUDE.md files are markdown documents used by Claude Code to store persistent project context, coding standards, and workflows so the AI understands your specific environment.

How does the token budget analysis work?

It estimates the token count of 'always-loaded' memory files and compares them against repository-specific thresholds, marking them as PASS, WARN, or FAIL based on size.

Can I audit only specific parts of my memory?

Yes, the skill supports arguments like 'project' or 'user' to limit the scope of the audit, or '--token-budget' to run only the size analysis without a full syntax audit.

Why should I audit my memory files?

Auditing prevents circular imports, ensures syntax correctness, and keeps your context within token limits, which prevents Claude from becoming confused or slow due to context bloat.

Claude Memory Auditor

Name: Claude Memory Auditor
Author: melodic-software

bymelodic-software

•

開発者ツール

Audits and optimizes CLAUDE.md memory files to ensure high-quality context, proper hierarchy, and efficient token usage.

The audit-memory skill is a specialized diagnostic tool designed to maintain the integrity and performance of Claude Code's memory system. It systematically scans CLAUDE.md files across enterprise, project, and user scopes to validate import syntax, detect circular dependencies, and enforce organizational standards. By providing detailed token budget analysis and progressive disclosure recommendations, it helps developers keep their AI context concise and effective, preventing context bloat and ensuring Claude always operates with the most relevant project information.

主な機能

01Performs token budget analysis to prevent context overflow

02Generates detailed audit logs with actionable remediation steps

03Enforces memory hierarchy across enterprise, project, and user scopes

04Validates @path import syntax and detects circular dependencies

0538 GitHub stars

06Identifies large files suitable for hub-pattern splitting or on-demand loading

ユースケース

01Optimizing Claude's response speed by pruning unnecessary 'always-loaded' context

02Maintaining complex project memory structures in large-scale repositories

03Debugging broken context imports or conflicting memory instructions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add melodic-software/claude-code-plugins audit-memory

For use in Claude.ai and ChatGPT

Download Skill