Keeps your API keys out of LLM context by storing them in a local AES-256-GCM vault and resolving placeholders like {{OPENAI_KEY}} at the MCP boundary. Exposes two tools: enigmagent_resolve to decrypt secrets and enigmagent_list to show what's stored. Every secret is pinned to a domain, so the resolver only returns values when the origin matches. Works with Claude Desktop, Cursor, Continue.dev, Cline, and anything else that speaks MCP over stdio. Also runs in REST mode if you need it behind a proxy. Reach for this when you're building agents that call third-party APIs and you don't want tokens living in chat logs or conversation history.
claude mcp add --transport stdio agnuxo1-enigmagent-mcp uvx enigmagent-mcp