- Home
- Skills
- Kamalnrf
- Claude Plugins
- Skills Discovery
skills-discovery_skill
- TypeScript
389
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill kamalnrf/claude-plugins --skill skills-discovery- SKILL.md5.6 KB
Overview
This skill helps you discover, install, and manage Agent Skills from the claude-plugins.dev registry so your agent gains specialized capabilities. It prioritizes proactive discovery: ask whether a skill could handle a task before you start work. Use it when tasks involve specific technologies, file formats, or domains where expert patterns and tools matter.
How this skill works
The skill queries the public registry API to search indexed skill names, descriptions, and tags, returning paginated results with metadata (namespace, stars, installs, description). It helps construct effective queries, recommends top matches, and guides installation commands for the appropriate client and scope. Before installing, it summarizes how each candidate skill specifically helps your current task and requests confirmation.
When to use it
- Before starting non-trivial work that could benefit from specialized tooling or best practices
- When tasks involve particular technologies, frameworks, or file formats (e.g., Django, PDF, ML pipelines)
- If you notice your advice becoming generic and suspect an expert skill exists
- When the user explicitly asks to find, install, list, or uninstall skills
- When preparing testing, deployment, API design, or domain-specific workflows
Best practices
- Search proactively with 1–3 specific terms (technology + task often works best)
- Show only the top 3–5 most relevant skills and explain how each helps the exact user task
- Ask which client the user uses before attempting installation (defaults to claude-code)
- Prefer skills with higher installs and stars but explain trade-offs for niche tools
- Always ask for user confirmation before installing or changing scope
Example use cases
- User: 'Create a Django REST API' → search for django-rest skills, show top results, offer to install or proceed manually
- User: 'Find skills for Python packaging' → return 3 relevant skills with installs and how they speed up packaging tasks
- User wants project-scoped tooling → recommend --local installs and show appropriate client command
- User: 'List my installed skills' → run the management listing workflow and offer uninstall help
FAQ
Use the registry API endpoint with q, limit, and offset parameters to fetch results.
Which client should I pick?
Ask the user which environment they use; default is claude-code. Use the client flag for other CLIs or editors.