MCP server for dynamic AI agent skill management
Project description
skills-as-mcp
An MCP server for dynamic AI agent skill management. Install, search, and manage Agent Skills from any MCP client — Claude Code, Cursor, Simba, or any tool that speaks MCP.
What Makes This Different
Every existing solution assumes skills are local files you manually place. skills-as-mcp lets your AI agent install skills from URLs mid-conversation:
You: "install this skill https://example.com/weather/SKILL.md"
Agent: calls install_skill(url="https://example.com/weather/SKILL.md")
Agent: "Weather skill installed. You now have weather lookup capabilities."
You: "what's the weather in Tokyo?"
Agent: reads skill://weather → follows instructions → answers
Install
pip install skills-as-mcp
Quick Start
With Claude Code
Add to your MCP config (.claude/settings.json or ~/.claude.json):
{
"mcpServers": {
"skills-as-mcp": {
"command": "skills-as-mcp"
}
}
}
With Any MCP Client
# Run via stdio (default)
skills-as-mcp
# Run as HTTP server
skills-as-mcp --transport http --port 8000
Storage
Skills are stored in ~/.skills-as-mcp/ by default. Override with:
export SKILL_SHELF_DIR=/path/to/shelf
MCP Tools
| Tool | Description |
|---|---|
install_skill(url?, content?) |
Install from URL or raw SKILL.md content |
list_skills(enabled_only?) |
List all installed skills |
get_skill(name) |
Read full SKILL.md instructions |
search_skills(query) |
Search by name/description |
remove_skill(name) |
Uninstall a skill |
enable_skill(name) |
Re-enable a disabled skill |
disable_skill(name) |
Disable without removing |
MCP Resources
| URI | Description |
|---|---|
skill://index |
Index of all enabled skills |
skill://{name} |
Full SKILL.md content for a skill |
SKILL.md Format
Follows the agentskills.io open standard:
---
name: my-skill
description: What this skill does and when to use it
license: MIT
compatibility: Requires internet access
metadata:
author: your-name
version: "1.0"
---
# My Skill
Instructions the agent reads and follows...
Examples
The examples/ directory contains sample skills you can install to test the server:
| Skill | Description |
|---|---|
web-search |
Search the web and summarize results |
code-reviewer |
Review code changes for bugs and improvements |
daily-standup |
Prepare daily standup summaries from git activity |
Install a sample skill from the repo:
# Start the server
skills-as-mcp
# Then from any MCP client, install a sample skill:
install_skill(url="https://raw.githubusercontent.com/sumododda/skills-as-mcp/main/examples/web-search/SKILL.md")
install_skill(url="https://raw.githubusercontent.com/sumododda/skills-as-mcp/main/examples/code-reviewer/SKILL.md")
install_skill(url="https://raw.githubusercontent.com/sumododda/skills-as-mcp/main/examples/daily-standup/SKILL.md")
Or install from a local file:
install_skill(content=open("examples/web-search/SKILL.md").read())
Development
git clone https://github.com/sumo/skills-as-mcp.git
cd skills-as-mcp
pip install -e ".[dev]"
pytest -v
License
MIT
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file skills_as_mcp-0.2.0.tar.gz.
File metadata
- Download URL: skills_as_mcp-0.2.0.tar.gz
- Upload date:
- Size: 12.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
e06ae68662fc1285395c321587b785a604628108114aef37205ea4780f7467a2
|
|
| MD5 |
ea8086e32e641863253537406be876aa
|
|
| BLAKE2b-256 |
87ca35a7d31fb3a73bde71843c01f593cf4f39f461bb719c597edeb7ef072afd
|
File details
Details for the file skills_as_mcp-0.2.0-py3-none-any.whl.
File metadata
- Download URL: skills_as_mcp-0.2.0-py3-none-any.whl
- Upload date:
- Size: 8.5 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
001d26822ef9b8624c719e60995160678ac3ac92e3c1556017f7200cee91e325
|
|
| MD5 |
e0603a5340f15f95bdfda007cd16b768
|
|
| BLAKE2b-256 |
720435dd756e43a1fd33025d1b9566c5b170316103ed7fb83ed2d2fd8f7bfa77
|