How does the similarity detection work?

It uses a keyword-bag approach to compare names, domains, and descriptions, flagging assets with over 40% overlap to prevent redundant logic.

Does it overwrite my manual changes?

No, the skill identifies manually created assets and explicitly protects them, only suggesting updates for assets marked as 'generated' in their frontmatter.

What triggers a 'stale' status for an asset?

Staleness is detected when there is a mismatch between an asset's registered domains and the current codebase analysis, or when required tools/skills are missing.

Can I use this to find global skills?

Yes, it scans project-specific directories, user-level home directories (~/.claude/skills), and plugin folders simultaneously.

What is cross-type matching?

It checks for overlap between skills and agents to ensure your AI assistant has a clean separation of concerns across different types of assets.

Asset Discovery & Management

Name: Asset Discovery & Management
Author: Kiwi-Home

byKiwi-Home

0•

개발자 도구

Scans and manages AI coding assets across project, user, and plugin layers to prevent duplication and ensure workflow consistency.

The Asset Discovery skill provides a robust mechanism for scanning and indexing Claude Code assets across multiple environments. It identifies existing skills and agents in project, user, and plugin directories, extracting metadata to facilitate workspace organization. Using advanced similarity heuristics based on keyword-bag overlap, the skill detects potential naming collisions and redundant capabilities before they are created. It also tracks the provenance of assets, distinguishing between manually authored files and AI-generated ones to safely suggest updates for stale components without risking manual work.

주요 기능

010 GitHub stars

02Automated staleness detection for domains, tools, and skills

03Multi-layer directory scanning (Project, User, and Plugin levels)

04Keyword-bag similarity heuristics for intelligent duplicate detection

05Standardized metadata extraction from YAML frontmatter

06Provenance classification to protect manual assets from overwrites

사용 사례

01Auditing a development environment to identify and update out-of-date AI agents

02Managing large-scale AI coding workflows across global and local configuration layers

03Preventing redundant skill generation by checking for existing domain overlaps

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kiwi-home/ai-coding-resources asset-discovery

For use in Claude.ai and ChatGPT

Download Skill