aidemd-mcp/server

Overview

An MCP server that implements the AIDE spec methodology, enabling AI agents to work with progressive disclosure specifications alongside code. It provides a standardized interface for agents to read, write, and manage spec files that unfold detail incrementally. The server is written in TypeScript and designed to integrate into AI coding workflows.

Best for

Best for
Developers building AI-assisted coding workflows that follow the AIDE progressive disclosure spec methodology

Use cases

• Teaching AI agents to follow progressive disclosure specification patterns
• Managing AIDE spec files alongside code in agent-assisted development
• Integrating structured spec methodology into MCP-compatible AI coding tools

How to use

Install

npx @aidemd-mcp/server@latest init

Tools exposed

• aide_discover
• aide_read
• aide_scaffold
• aide_inspect
• aide_validate
• aide_info
• aide_upgrade
• aide_brain
• aide_sync

Tested with

Claude Code, Claude Desktop, Cursor, VS Code, Windsurf

Example client config

{\n  "mcpServers": {\n    "aide": {\n      "command": "npx",\n      "args": ["-y", "@aidemd-mcp/server@latest"]\n    }\n  }\n}

Notes

An MCP server that implements the AIDE spec methodology, enabling AI agents to work with progressive disclosure specifications alongside code. It provides a standardized interface for agents to read, write, and manage spec files that unfold detail incrementally. The server is written in TypeScript and designed to integrate into AI coding workflows.

2 stars on GitHub. Last updated 2026-05-22. Licensed MIT.

Use cases

• Teaching AI agents to follow progressive disclosure specification patterns
• Managing AIDE spec files alongside code in agent-assisted development
• Integrating structured spec methodology into MCP-compatible AI coding tools

Pros

• Enforces a structured spec methodology that reduces ambiguity for AI agents
• Open source with TypeScript implementation, easy to extend or audit
• Lightweight MCP server, straightforward to self-host

Cons

• Very early stage with only 2 GitHub stars, limited community or support
• Narrowly scoped to the AIDE spec methodology, not general-purpose
• Requires AI agents to be MCP-aware, which is still a nascent ecosystem

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.