AI agents call tor_private_fetch to retrieve information from Tor without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool fetches web content (Read category). However, severity is elevated to medium because: (1) the description is empty, reducing confidence; (2) fetching through Tor can be used to access sensitive/restricted content or bypass security controls; (3) the sibling tools (tor_new_identity, tor_privacy_status) suggest active adversarial/evasion capabilities.
From the tool's definition Tool named 'tor_private_fetch' on a server that 'routes web requests...for anonymized fetching'; the function name and server purpose indicate data retrieval via HTTP/HTTPS requests through Tor.
Attacks that exploit this kind of access
tor_private_fetch. It is categorised as a Read tool in the Tor MCP Server, which means it retrieves data without modifying state.
Register the Tor MCP server in PolicyLayer and add a rule for tor_private_fetch: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Tor. Nothing to install.
tor_private_fetch is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the tor_private_fetch rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for tor_private_fetch. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
tor_private_fetch is provided by the Tor MCP server (rushikeshmore/tor-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →