@morphism-systems/agentic-math

Category theory MCP server for AI agents. Morphism composition, functor mapping, natural transformations, entropy computation, and convergence checking.

About

A standalone MCP server that exposes category theory operations to AI assistants. Compose morphisms, map functors, check naturality conditions, compute Shannon entropy, and verify convergence -- all through structured tool calls. Free to use with any MCP-compatible client.

Installation

npx @morphism-systems/agentic-math

Quick Start

Add to your .claude/mcp.json or Cursor MCP settings:

{
  "mcpServers": {
    "morphism-math": {
      "command": "npx",
      "args": ["@morphism-systems/agentic-math"]
    }
  }
}

Tools

| Tool | Description | |------|-------------| | morphism_compose | Compose two morphisms f . g | | functor_map_object | Apply a functor to an object | | functor_map_morphism | Apply a functor to a morphism | | natural_transform_check | Check naturality condition | | natural_transform_compose | Vertical composition of natural transformations | | entropy_compute | Shannon entropy H(s) = -sum(p_i ln(p_i)) | | convergence_check | Check if governance converges (kappa < 1) | | drift_check | Compute drift delta between states | | analyze_scores | Full governance score analysis |

Example Output

// Tool: morphism_compose
// Input: { "f": { "name": "g", "source": "B", "target": "C" },
//          "g": { "name": "f", "source": "A", "target": "B" } }
// Output:
{ "name": "g . f", "source": "A", "target": "C" }

// Tool: entropy_compute
// Input: { "values": [0.9, 0.8, 1.0, 0.7, 1.0, 0.95, 0.85] }
// Output:
{ "entropy": 0.42, "normalized": true }

Documentation

Full reference: https://morphism-systems.github.io/morphism/

License

MIT