ExpertPack MCP
ExpertPack MCP serves structured knowledge packs to AI agents via stdio or HTTP, enabling hybrid search, graph traversal, and provenance-grounded retrieval. It supports multi-pack hosting, MCP Skills, and freshness tracking to provide verifiable, token-efficient context for domain-specific queries.
Package and access
Local package with source-attributed installation metadata. Koskamo does not execute packages during review.
What you can do with it
- Serve domain-specific knowledge packs to AI agents
- Perform hybrid search with provenance and graph traversal
- Expose knowledge packs as MCP Skills
- Validate pack structure and freshness
Source and maintenance
- Maintenance checked
Packages and installation
Source-provided package data. Koskamo has not executed or live-probed this package.
ep-mcp 0.7.0
{
"version": "0.7.0",
"transport": "stdio",
"identifier": "ep-mcp",
"runtimeHint": "uvx",
"registryType": "pypi",
"registryBaseUrl": "https://pypi.org",
"packageArguments": [
{
"type": "positional",
"value": "serve"
},
{
"name": "--config",
"type": "named",
"format": "filepath",
"isRequired": true,
"description": "Path to an ep-mcp config YAML listing your ExpertPacks"
},
{
"name": "--transport",
"type": "named",
"value": "stdio"
}
],
"environmentVariables": [
{
"name": "GEMINI_API_KEY",
"description": "Gemini API key for embeddings (or set AZURE_OPENAI_API_KEY when using Azure OpenAI)"
},
{
"name": "AZURE_OPENAI_API_KEY",
"description": "Azure OpenAI key, if the config uses the Azure embedding provider"
}
]
}Tools
Reported by the attributed source; not independently observed by Koskamo.
Connect from your client
Each snippet configures the same source-provided package; replace placeholders before use.
Cursor
{
"mcpServers": {
"expertpack-mcp": {
"command": "uvx",
"args": [
"ep-mcp==0.7.0",
"serve",
"--transport=stdio"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"AZURE_OPENAI_API_KEY": "<YOUR_AZURE_OPENAI_API_KEY>"
}
}
}
}Cline
{
"mcpServers": {
"expertpack-mcp": {
"command": "uvx",
"args": [
"ep-mcp==0.7.0",
"serve",
"--transport=stdio"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"AZURE_OPENAI_API_KEY": "<YOUR_AZURE_OPENAI_API_KEY>"
}
}
}
}VS Code
{
"servers": {
"expertpack-mcp": {
"type": "stdio",
"command": "uvx",
"args": [
"ep-mcp==0.7.0",
"serve",
"--transport=stdio"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"AZURE_OPENAI_API_KEY": "<YOUR_AZURE_OPENAI_API_KEY>"
}
}
}
}Claude Desktop
{
"mcpServers": {
"expertpack-mcp": {
"command": "uvx",
"args": [
"ep-mcp==0.7.0",
"serve",
"--transport=stdio"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"AZURE_OPENAI_API_KEY": "<YOUR_AZURE_OPENAI_API_KEY>"
}
}
}
}Let your agent find its tools
Find servers like ExpertPack MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.