skills-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@skills-mcp what skills are available?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
skills-mcp
MCP server that serves agent skills (SKILL.md bundles, the format used by
Claude Code plugin skill repos like cloudflare/skills)
to any MCP client — including clients that can't install plugins, such as
claude.ai / Claude Desktop connectors or an aggregator like MetaMCP.
Skills are exposed two ways:
Tools —
list_skills,read_skill,read_skill_filefor tool-oriented clients. The server's MCPinstructionslist every skill so models know when to load one.Resources —
skill://<name>/SKILL.md(plus each bundled file) for resource-aware clients.
Configuration
Env var | Meaning |
| Comma-separated skill sources (required). Each entry is |
| Cache directory for GitHub tarballs (default: OS tmpdir). |
GitHub sources are fetched as tarballs on startup (no git needed) and fall
back to the cached copy if the fetch fails.
Related MCP server: Agent Skills MCP
Usage
SKILLS_SOURCES=cloudflare/skills,raymatos/my-skills npx -y skills-mcp-serverAny directory containing a SKILL.md (searched up to 6 levels deep) is a
skill; YAML frontmatter name: and description: are used when present.
License
MIT
Available Tools
3 toolslist_skillsList SkillsARead-onlyIdempotent
List all available agent skills with name, description, source, and bundled files. Call read_skill to load one.
Inputs: no named parameters.
read_skillRead SkillARead-onlyIdempotent
Read a skill's SKILL.md instructions. Use the returned guidance for the current task; it may reference supporting files readable via read_skill_file.
Inputs: name (required).
read_skill_fileRead Skill FileARead-onlyIdempotent
Read a supporting file bundled with a skill (paths are listed by list_skills and referenced from SKILL.md).
Inputs: name (required), path (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 3 tools
Each tool has a distinct, non-overlapping role: listing skills, reading the main skill instructions, and reading supporting files. There is no ambiguity about which tool to use for a given step.
All tool names follow a consistent verb_noun pattern: list_skills, read_skill, read_skill_file. The verb and noun are clear and predictable.
Three tools is perfectly scoped for a skills management server. Each tool is necessary and there are no redundant or extraneous tools.
The tool surface covers the full workflow for accessing skills: discovery, reading the primary instructions, and accessing bundled files. No obvious gaps for the stated purpose.
Maintenance
Related MCP Connectors
Search and discover Agent Skills from the skills.sh registry. Powered by HAPI MCP server.
Hosted MCP server for live public-data APIs and Skills for AI agents.
Agent communication platform for agent to agent messaging via MCP. Messages, channels, skills.
Read-only MCP server for the Durable Agent Directory: search and compare agent-infrastructure tools.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceDual-interface MCP server that exposes declarative tools and reusable skills via standard MCP protocol and provides a REST API for authoring them.Apache 2.0
- AlicenseNot gradedqualityBmaintenanceAn MCP server that makes Agent Skills available to any MCP-compatible agent through a declarative, package.json-based configuration, enabling team-shareable skill management and execution.5MIT
- AlicenseNot gradedqualityBmaintenanceMCP server that unifies and synchronizes AI coding skills across multiple tools, exposing skill discovery and retrieval via list_skills and read_skill.347 npm14MIT
- AlicenseNot gradedqualityCmaintenanceServes Agent Skills over MCP using the SEP-2640 Skills Extension, enabling skills discovery, retrieval, and resource reading via MCP.Apache 2.0