MCP Skills for Claude Code, Codex & AI Agents

These skills help an agent expose tools and resources through MCP or operate an existing MCP system. Members cover server builders, OpenAPI conversion, WebMCP, context stores, vision tools, and MCP diagnostics.

25 skills

Explore MCP skills

3 skills
ModLens Vision Skill — Text-only DeepSeek reading a tweet screenshot in full detail via ModLens
Free
Other utilities

ModLens Vision Skill

Give text-only agents image understanding through a configurable vision CLI, covering image questions, screenshots, and chart interpretation.

4.9KStars154Forks
Repository last pushed
Agent Vision Toolkit — HTML and CSS reconstruction of the model-training infographic
Free
Other utilities

Agent Vision Toolkit

Equip text-only agents with image questions, OCR, target localization, cropping, SVG tracing, and screenshot comparison tools.

1.3KStars50Forks
Repository last pushed
ModSearch Web Research — Text-only DeepSeek summarising a blog link through ModSearch
Free
Research & data

ModSearch Web Research

Add web search, X search, and page retrieval to agents through configurable search engines and a reusable command-line research workflow.

976Stars40Forks
Repository last pushed

Related topics

How to choose MCP skills

Use a builder when exposing a new capability, OpenAPI conversion for an existing contract, and context or diagnostic entries when the MCP system already runs.

  • Build an MCP server for an internal API
  • Convert an OpenAPI operation into an MCP tool
  • Inspect context usage in an MCP workflow