- Home
- MCP servers
- Skill Jack
Skill Jack
- typescript
0
GitHub Stars
typescript
Language
6 months ago
First Indexed
3 months ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
Installation
Add the following to your MCP client configuration file.
Configuration
View docs{
"mcpServers": {
"olaservo-skill-jack-mcp": {
"command": "skill-jack-mcp",
"args": [],
"env": {
"SKILLS_DIR": "YOUR_SKILLS_DIR"
}
}
}
}Skill Jack MCP is an MCP server that injects dynamic Agent Skill metadata into your client’s context and serves on-demand Skill content to your LLM. It lets you discover skills from your MCP workspace roots or a fallback directory, provide real-time updates, and load full Skill content when needed. This enables your LLM to reason with up-to-date skills and execute them through a structured tool flow.
How to use
With MCP Roots (recommended) your client shares workspace roots and Skill discovery happens automatically. The server scans each root for Skill directories and makes metadata available to the LLM, which can then request full Skill content on demand.
Without Roots support or when you want a default location, provide a Skill directory. The server will use that directory to load and serve Skills and their resources.
How to install
Prerequisites: ensure you have Node.js installed (LTS version recommended). You do not need additional runtimes.
# Install dependencies
npm install
# Build the server
npm run build
Run the server with Roots enabled (default behavior): the server will discover skills from your client workspace automatically.
skill-jack-mcp
If you do not have Roots support or want to point to a specific skills directory, pass the path or set an environment variable.
skill-jack-mcp /path/to/skills
or
SKILLS_DIR=/path/to/skills skill-jack-mcp
Windows users: when using MCP Inspector, use forward slashes in paths.
skill-jack-mcp "C:/Users/you/skills"
Notes on usage and behavior
On connection, Skill Jack MCP asks the client for workspace roots (or uses the fallback directory). It then discovers Skills under each root and injects Skill metadata into the system prompt as available_skills. When you request a full Skill, the server returns the complete SKILL.md content. Real-time updates are supported when workspace roots change, ensuring your Skill set stays current.
Security considerations
Skills are treated as trusted content. The server serves Skill files directly to clients without sanitization. Only load Skills from directories you trust. Path traversal is prevented, and there are safeguards on file size and directory depth. Remember that Skill content can influence LLM behavior by design.
Available tools
skill
Load and activate an Agent Skill by name. Returns the full SKILL.md content.
skill-resource
Read files within a skill directory (scripts/, references/, assets/, snippets/), supporting progressive disclosure and file listing.