Runtime trust layer for MCP tools. Detects post-approval drift including behavioral effective-permission drift (denied→allowed) that manifest diffing misses and quarantines before execution, and emits tamper-evident Security Receipts.

Open-source, local-first workspace for you and your AI agents: notes, tasks, memory, and a coding harness.

MCP server providing Quasar Framework documentation access for Claude Code and other MCP clients

📚 Access Quasar Framework documentation seamlessly with this MCP server, featuring full-text search, component details, and smart caching for efficiency.

MCP server for Cortex AI — exposes RAG pipeline as 3 tools to Claude Desktop and Cursor via Model Context Protocol

GitHub MCP agent using Haystack + Gemini to list and summarize issues, PRs, and commits (Docker stdio or HTTP).