discovering-skills

Finds existing skills before building new ones. Searches local installations, official Anthropic repository, and community collections. Evaluates fit and recommends installation or adaptation.

⭐0stars🔀0forks📁View on GitHub🕐Updated Jan 9, 2026⬇️4downloads

When & Why to Use This Skill

The Discovering Skills Claude skill optimizes agent development by automatically searching local, official, and community repositories for existing capabilities. It streamlines the workflow for developers by identifying pre-built tools, evaluating their fit for specific requirements, and recommending whether to install, adapt, or create new skills, thereby preventing redundant coding and accelerating deployment.

Use Cases

  • Preventing Redundant Development: Before writing a new tool from scratch, use this skill to check if a similar capability already exists in the official Anthropic repository or community collections.
  • Local Skill Management: Quickly scan local directories like ~/.claude/skills/ to identify and reuse previously developed or installed tools for new agent projects.
  • Community Resource Integration: Leverage the 'awesome-claude-skills' ecosystem by searching community-vetted repositories to find specialized functionalities that enhance agent performance.
  • Gap Analysis and Adaptation: Evaluate how well an existing skill fits a new use case and receive recommendations on whether to use it as-is or adapt it using the 'adapting-skills' workflow.
namediscovering-skills
descriptionFinds existing skills before building new ones. Searches local installations, official Anthropic repository, and community collections. Evaluates fit and recommends installation or adaptation.
allowed-toolsWebSearch, WebFetch, Read, Glob

Discovering Skills

Before creating a skill, search for existing ones.

Search Order

  1. Local - ~/.claude/skills/, .claude/skills/
  2. Official - github.com/anthropics/skills
  3. Community - awesome-claude-skills repositories

Process

  1. Identify the capability needed
  2. Search sources in order
  3. Evaluate fit
  4. Recommend: install, adapt, or create new

When Found

Report: skill name, source, what it does, how well it fits, install command.

When Not Found

Report what was searched and offer: create new, adapt similar, or proceed without.

Reference Documentation

Integration

  • Imperfect fit → suggest adapting-skills
  • Not found → suggest detecting-skill-gaps
  • Multiple found → let user choose best fit