io.github.swnotmetal/koma-core-mcp — Model Context Protocol (MCP) Server

This MCP server exposes Koma Core’s protected RAG storage to AI agents as MCP tools. Agents can discover public metadata via a search index, while full document content is gated behind access tiers using an access token.

🛠️ Key Features

• Protected RAG storage for agent access
• Public index discovery without returning content
• Token- and tier-based authorization for full content retrieval
• Includes tools: search_docs and retrieve_doc

🚀 Use Cases

• Prevent “discovery ≠ authorization” mistakes by separating metadata search from private retrieval
• Guard agent workflows where private RAG content must be access-controlled

⚡ Developer Benefits

• search_docs: returns metadata, never content (Public access)
• retrieve_doc: fetches full content by source ID with userTier gating (Requires userTier ≥ document tier)

⚠️ Limitations

• Full document content retrieval depends on correct access tier (userTier requirement)

Topics

ai-security, guardrails, middleware, nodejs, prompt-injection, rag, rate-limiting, typescript, ai-firewall, llm-security, owasp-top-10-llm, ai-governance, ai-tools