Skip to main content
Glama

MCP Spec Generator

by Huxley-Brown

MCP Spec Generator (docs)

Minimal notes for running and developing the MCP spec-generator tool described in this repo.

Requirements

  • Node.js LTS (>=18)
  • Environment variable: ANTHROPIC_API_KEY (store securely)

Quick notes

  • The server communicates with Anthropic's claude-4-sonnet-20240620 model. Keep temperature at 0.0 for deterministic output.
  • Do not run with write permissions in production: MCP tools should return create_file/edit_file calls for user approval in Cursor.
  • Example .cursor/mcp.json is provided in the repo as a template. When configuring, whitelist only the project folder.

Testing

  • There are fixtures in tests/fixtures/ to validate the expected Markdown outputs from the model.

How to run

  • Run tests: node tests/run_tests.js or npm test.
  • Dry-run (propose files): node index.js "My project idea..."
  • Persist files: node index.js "My project idea..." --allow-write
  • Read prompt from stdin: echo "My idea" | node index.js

Linting & formatting

  • ESLint and Prettier configs are included. To lint/format locally run:
    • npm run lint
    • npm run format

Environment & setup (required)

  • Set your Anthropic API key in the environment before running the tool:
    • Linux/macOS: export ANTHROPIC_API_KEY=sk-...
    • Windows (PowerShell): setx ANTHROPIC_API_KEY "sk-..."
  • Optionally copy env.example to .env or export additional overrides like MCP_AUDIT_FILE or ANTHROPIC_TIMEOUT_MS.
  • Make the CLI executable if you prefer: chmod +x index.js

Cursor MCP configuration

  • The repo includes .cursor/mcp.json configured to whitelist the project path. Verify it points to your repository root before registering with Cursor.

Security

  • Never point the Filesystem MCP server at ~ or your home directory. Whitelist only the necessary project paths.
-
security - not tested
F
license - not found
-
quality - not tested

hybrid server

The server is able to function both locally and remotely, depending on the configuration or use case.

Generates project specifications and file structures using Claude AI through MCP integration. Enables users to describe project ideas and automatically create corresponding documentation and project files.

Related MCP Servers

  • -
    security
    A
    license
    -
    quality
    AI-powered MCP server that enables Claude and other LLMs to interact directly with construction documents, drawings, and specifications through advanced RAG and hybrid search capabilities.
    Last updated -
    3
    TypeScript
    MIT License
  • -
    security
    F
    license
    -
    quality
    A comprehensive, domain-agnostic documentation scraping and AI integration toolkit. Scrape any documentation website, create structured databases, and integrate with Claude Desktop via MCP (Model Context Protocol) for seamless AI-powered documentation assistance.
    Last updated -
    5
    Python
    • Apple
  • -
    security
    A
    license
    -
    quality
    An MCP server that automatically manages and organizes project documentation using the document reference pattern, keeping CLAUDE.md files clean and under 500 lines while maintaining full context for AI assistants.
    Last updated -
    Python
    MIT License
  • A
    security
    A
    license
    A
    quality
    Connects Blender to Claude AI through the Model Context Protocol (MCP), enabling prompt-assisted 3D modeling, scene creation, and manipulation.
    Last updated -
    17
    Python
    MIT License
    • Apple

View all related MCP servers

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Huxley-Brown/Project-Setup-MCP'

If you have feedback or need assistance with the MCP directory API, please join our Discord server