"Roots" matching MCP servers:

Enables AI agents to inspect, draft, validate, and locally verify contract changes against frozen policies, with separate input and output roots and no approval or verified-signing capability.

Enforces authorized file access through MCP roots, blocking path traversal, symlink escapes, and absolute path escapes before any read.

Enables self-hosted remote access to a user-controlled device's approved filesystem roots, allowing clients to list directories, read and search files, inspect metadata, ping the agent, and list devices through an MCP tunnel.

Enables controlled local Windows file inspection, reading, writing, reversible deletion, restoration, and auditing across user-selected folder roots. Commits occur only after separate owner approval in a native desktop dialog, while the tool cannot bypass OS security controls.

Enables MCP clients like ChatGPT to securely access and manage a controlled local directory over a remote, OAuth-protected endpoint, with configurable filesystem roots and ngrok tunneling.

Enables deploying a secure MCP gateway that proxies, sandboxes, and policy-governs third-party MCP servers, with credential vaulting and audited tool access across stdio, SSE, and Streamable HTTP.

Enables multi-user, per-namespace storage of AES-256-GCM-encrypted credentials (SSH, web, API, DB) managed through a web console with TOTP 2FA, and lets AI clients save, list, retrieve, or server-side invoke those secrets over a remote Streamable-HTTP endpoint without ever exposing the values in chat.