SkillMesh MCP Server

SkillMesh

Stop stuffing hundreds of tools into your LLM prompt. Route to the right ones.

SkillMesh is a retrieval router for agent tool catalogs. Instead of loading every skill/tool into every prompt, it selects the best few cards for the query and injects only those.

Why Teams Adopt SkillMesh

• Keeps prompts small as your catalog grows (top-K instead of full dump)
• Improves tool selection quality on multi-domain tasks
• Cuts token cost per call by avoiding irrelevant tool context
• Works with Claude (MCP), Codex (skill bundle), and local CLI workflows
• Standardized OpenAI-style function schemas for tool expansion

The Problem

LLM agents break when you load every tool into the prompt. Token counts explode, accuracy drops, and cost scales linearly with your catalog size. Teams with 50+ skills end up with bloated system prompts that confuse the model and burn budget.

SkillMesh solves this with retrieval-based routing: given a user query, it selects only the top-K most relevant expert cards and injects them into the prompt — keeping context small, accurate, and cheap.

High-Value Use Cases

• Internal AI assistants with large tool/skill catalogs (50+ cards)
• Multi-step workflows crossing domains (data -> ML -> infra -> reporting)
• Teams using MCP where tool overload hurts selection quality
• Role-based execution flows (Data-Analyst, Financial-Analyst, AWS-Engineer)

SkillMesh vs Static Skill Docs

Before vs After

How It Works

User Query
    │
    ▼
┌─────────────────────┐
│  BM25 + Dense Index  │  ← Scores every card in your registry
└─────────┬───────────┘
          │
          ▼
┌─────────────────────┐
│   RRF Fusion Rank    │  ← Merges sparse + dense rankings
└─────────┬───────────┘
          │
          ▼
┌─────────────────────┐
│   Top-K Card Select  │  ← Returns the K best expert cards
└─────────┬───────────┘
          │
          ▼
┌─────────────────────┐
│  Agent acts as expert │  ← Full instructions injected into prompt
└─────────────────────┘

Each card contains: execution behavior, decision trees, anti-patterns, output contracts, and composability hints — everything the agent needs to act as a domain expert.

One-line MCP install (Claude Desktop / Claude Code)

Add this to your Claude Desktop config (claude_desktop_config.json) or Claude Code MCP settings:

{
  "mcpServers": {
    "skillmesh": {
      "command": "uvx",
      "args": ["--from", "skillmesh[mcp]", "skillmesh-mcp"]
    }
  }
}

No env vars. No file paths. No cloning. The bundled registry is included in the package.

Requires uv to be installed.

60-Second Demo

git clone https://github.com/varunreddy/SkillMesh.git
cd SkillMesh
pip install -e .
skillmesh emit \
  --provider claude \
  --registry examples/registry/tools.json \
  --query "clean messy sales data, train a baseline model, and generate charts" \
  --top-k 5

Output (truncated):

<context>
  <card id="data.data-cleaning" title="Data Cleaning and Validation Expert">
    # Data Cleaning and Validation Expert
    Specialist in detecting and correcting data quality issues...
  </card>
  <card id="ml.sklearn-modeling" title="Scikit-learn Modeling and Evaluation">
    ...
  </card>
  <card id="viz.matplotlib-seaborn" title="Visualization with Matplotlib and Seaborn">
    ...
  </card>
</context>

Only the relevant experts are injected — the rest of the 100+ card catalog stays out of the prompt.

Integrations