MCP Servers

A collection of Model Context Protocol servers, templates, tools and more.

A unified Model Context Protocol (MCP) server that aggregates, indexes, and dynamically serves over 200+ specialized Agent Skills, Prompts, and Coding Workflows to AI assistants.

Created 2/22/2026
Updated 2 days ago
Repository documentation and setup instructions

EVOKORE-MCP v2.0 (The Enterprise Router)

A unified Model Context Protocol (MCP) server that aggregates, indexes, and dynamically serves over 200+ specialized Agent Skills, while functioning as a Multi-Server Aggregator. EVOKORE-MCP v2.0 proxies traffic to child servers (e.g., GitHub, Filesystem), namespaces their tools dynamically, and governs execution with a Human-in-the-Loop (HITL) Security Interceptor.

🚀 Features

  • Multi-Server Aggregation (Proxy Layer): Bootstraps child MCP servers from mcp.config.json and routes JSON-RPC messages via stdio, acting as a single unified endpoint for AI clients.
  • Dynamic Tool Prefixing: Automatically namespaces proxied tools (e.g., fs_read_file, github_create_issue) to prevent collisions and keep contexts organized.
  • Human-in-the-Loop (HITL) Security Interceptor: A stateless security architecture that intercepts restricted tool calls and enforces explicit human approval via the _evokore_approval_token, protecting sensitive endpoints.
  • Active Skill Orchestration: Built-in skills (like docs_architect and skill_creator) are directly integrated into the SkillManager, allowing EVOKORE to natively harness child server tools (like reading the filesystem) before returning contextualized prompts to the AI.
  • Dynamic Skill Loading: Automatically scans the SKILLS/ directory on startup. Keeps context windows lean by using lightweight semantic search (fuse.js) to serve prompts only when requested via the resolve_workflow tool.
  • Massive Built-in Library: Includes 146 specialized developer plugins (Wshobson), the Hive Framework, the Manus planning-with-files protocol, and extensive custom orchestration workflows (e.g., arch-aep-runner, session-wrap).

🛠️ Setup

To use this server with your AI client (like Claude Desktop or Cursor), add the following to your MCP configuration file:

{
  "mcpServers": {
    "evokore-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/EVOKORE-MCP/dist/index.js"]
    }
  }
}

🎓 Comprehensive Training

Dive into our extensive, deeply researched use cases and training guides for all 200+ skills: Training & Use Cases Documentation.

📂 Repository Structure

  • src/index.ts: The core MCP Server implementation.
  • SKILLS/: The library of 200+ organized Markdown skills.
  • scripts/: Maintenance scripts (like clean_skills.js) to normalize YAML frontmatter across the library.

🤝 Contributing

This repository uses a strict Pull Request (PR) only workflow. Direct commits to the main branch are restricted (except by the owner).

  1. Fork or branch from main.
  2. Add or update a skill in the SKILLS/ directory. Ensure it has valid YAML frontmatter.
  3. Submit a Pull Request.
Quick Setup
Installation guide for this server

Install Package (if required)

uvx evokore-mcp

Cursor configuration (mcp.json)

{ "mcpServers": { "mattmre-evokore-mcp": { "command": "uvx", "args": [ "evokore-mcp" ] } } }