This is the MCP component of envibe, a permission layer that sits between AI assistants and your .env files. It exposes five tools: env_list shows variables with their access levels, env_get and env_set handle reads and writes based on permissions, env_describe returns format hints and examples, and env_check_required validates setup. You configure access levels (full, read-only, placeholder, schema-only, hidden) in a manifest file, and the server automatically generates a filtered .env.ai view while blocking direct access to your actual secrets. Useful when you want Claude to help with configuration without exposing production credentials or API keys.
claude mcp add --transport stdio io.github.dominic-righthere-envibe -- npx -y envibe-mcp