skills_find_relevantLABSCO SUMMARY
Install the master skill file or the workflow does not fire.
The three-tier discipline — search, load, supplement — depends on a master skill file dropped into the agent's project root. Without it, the README says agents skip the score thresholds, load skill bodies speculatively and ignore the tier-3 manifest, which burns exactly the context the design was meant to save. Two more constraints worth planning for: script execution only works against the local Python server, and search quality is only as good as the trigger phrases written into each SKILL.md — one vague skill degrades results across the registry.
What it is
A registry that serves Agent Skills — the SKILL.md format — over MCP instead of copying skill files into every repo and tool. An agent searches the registry by what it is trying to do, reads the confidence scores, and loads the full playbook for the one that matches.
What you get
- Seven tools in three tiers —
skills_find_relevant(query, top_k)for semantic search andskills_list_all(limit, offset)for browsing skills_get_body(skill_id, version?)returns the full instructions in one call, plus a manifest of what else the skill carries;versionpins to a specific releaseskills_get_options(skill_id)returns config schema, variants, dependencies and limitationsskills_get_reference(skill_id, filename)andskills_get_asset(skill_id, filename)fetch supporting docs and templates only when the instructions name themskills_run_script(skill_id, filename, input_data)executes a bundled script — local server only- 30+ bundled skills distilled from official documentation, each carrying complexity level, time estimate, prerequisites, use cases and a source URL
- Only the frontmatter is embedded — description plus trigger phrases, around 100 tokens — which keeps the search space clean of instruction prose
Requirements
You deploy your own. It needs a free Qdrant Cloud cluster (URL and API key) and a Cloudflare account — the Worker runs on the free plan, which covers 100k requests/day, because it uses SQLite-backed Durable Objects. Python 3.11+ runs the seeding script and the optional local server; Node.js 18+ provides the wrangler CLI. Published as skill-mcp version 1.2.0. Deploy with the one-click button or bash scripts/setup.sh, then seed Qdrant with python -X utf8 -m skill_mcp.seed.seed_skills using QDRANT_URL, QDRANT_API_KEY, WORKERS_AI_ACCOUNT_ID and WORKERS_AI_API_TOKEN. Clients connect over SSE to https://skill-mcp.<your-subdomain>.workers.dev/sse; skills_run_script needs the local Python server instead, since Workers cannot spawn subprocesses.
Setup effort
Build from source — git clone https://github.com/Jignesh-Ponamwar/skills-mcp.git
USE CASES
01
Getting an agent the playbook for an unfamiliar integration
Search by what you are building and load the matching skill's full instructions.
02
Sharing procedures across tools and teammates
Keep one registry every agent reads at runtime instead of copying skill files per repo.
03
Keeping context usage down on a long task
Load the body first, and fetch reference docs only when the instructions name them.
04
Pinning behaviour to a known-good version
Request a specific release of a skill so an update does not change what the agent does.
05
Seeing what a skill needs before using it
Read its config schema, dependencies and stated limitations first.
7 tools
Listed in project docs. Your client may show a different list.
skills_list_allskills_get_bodyskills_get_optionsskills_get_referenceskills_run_scriptskills_get_assetBEFORE YOU INSTALL
What to know first
Tools you get
7 tools named in the project's documentation.
Requires
- A Qdrant Cloud cluster — its URL and API key (free tier)
- A Cloudflare account — account ID and a Workers AI API token (Workers Free plan)
- Python 3.11+ and Node.js 18+ for the setup wizard and wrangler
INSTALLATION
Connect your client
This one is built from source rather than installed with a single command.
Server setup
TERMINAL
git clone https://github.com/Jignesh-Ponamwar/skills-mcp.git
cd skills-mcpRun the setup wizard from the project root. It creates .env and waits while you enter QDRANT_URL, QDRANT_API_KEY, WORKERS_AI_ACCOUNT_ID and WORKERS_AI_API_TOKEN, then installs the Python dependencies, seeds Qdrant with the bundled skills, deploys the Worker with wrangler and pushes the secrets:
TERMINAL
bash scripts/setup.shMCP Clients
Claude Code
CLAUDE CODE · JSON
{
"mcpServers": {
"skill-mcp": {
"type": "sse",
"url": "https://skill-mcp.<YOUR_SUBDOMAIN>.workers.dev/sse"
}
}
}The four credentials go into the .env file the setup wizard creates, not into the client block. <YOUR_SUBDOMAIN> is your Cloudflare workers.dev subdomain — the Worker address printed when the deploy finishes.
Cursor
CURSOR · JSON
{
"mcpServers": {
"skill-mcp": {
"url": "https://skill-mcp.<YOUR_SUBDOMAIN>.workers.dev/sse"
}
}
}The four credentials go into the .env file the setup wizard creates, not into the client block. <YOUR_SUBDOMAIN> is your Cloudflare workers.dev subdomain — the Worker address printed when the deploy finishes.
