What is the difference between renaming and canonicalization?

Renaming completely replaces the old tag with a new one, while canonicalization sets a primary name for a group of tags, often treating the others as aliases.

Can I target specific parts of my memory for cleanup?

Yes, you can specify a 'domain' argument to limit the analysis to a specific area, or choose to analyze your entire memory history.

Does this skill automatically change my tags?

No, the skill generates a detailed proposal first and only executes the specific actions (rename or canonicalize) that you explicitly approve.

How does the skill determine which tags are related?

It uses co-occurrence analysis and Pointwise Mutual Information (PMI) to calculate the statistical strength of the relationship between tags used together.

Where can I see the history of my tag changes?

Every time a cleanup is performed, the results are saved as a 'material' entry in your memory with the title 'Tag整理結果' and a date stamp.

Tag Cleanup & Optimization

Name: Tag Cleanup & Optimization
Author: isizono

byisizono

0•

コンテンツ管理

Analyzes tag co-occurrence to suggest and execute organizational improvements for your Claude Code memory system.

The Tag Cleanup skill helps maintain a clean and structured knowledge base within Claude Code's memory system by performing sophisticated co-occurrence analysis. It identifies duplicate tags, creates clusters of related topics, and spots 'orphan' tags that should be merged or deleted. The skill provides a structured proposal including suspected duplicates and clusters, allowing you to approve specific renames or canonicalizations. This ensures your memory remains searchable and organized without manual overhead, while keeping you in control through an approval-based execution workflow.

主な機能

01Automated tag co-occurrence analysis using Pointwise Mutual Information (PMI)

02Discovery of tag clusters for potential group consolidation or new parent tags

03Detection of low-usage orphan tags with suggested merge targets

040 GitHub stars

05Comprehensive logging of all cleanup actions saved directly to memory

06Identification of suspected duplicate tags based on naming similarity

ユースケース

01Reducing memory clutter by merging rarely used tags into more frequent, relevant categories

02Consolidating similar tags like 'checkin' and 'check-in' to improve search accuracy

03Standardizing naming conventions across a large body of documented work items

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add isizono/cc-memory tag-cleanup

For use in Claude.ai and ChatGPT