AI agents call ai.extract to retrieve information from Mcp without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and parses content from a URL and structures it according to a schema, but does not modify, delete, or execute code. The operation is purely informational (Read category). Severity is low because the primary risk is exposure of sensitive data at the fetched URL or misuse of the schema definition; the tool itself has no destructive, financial, or code-execution capabilities.
From the tool's definition Tool description states 'Fetch a URL and extract typed data from its content' — fetching and extracting are read-only operations with no modification, deletion, or code execution.
Attacks that exploit this kind of access
Fetch a URL and extract typed data from its content per a user-supplied JSON Schema. Use when you need a structured payload conforming to your own shape. It is categorised as a Read tool in the Mcp MCP Server, which means it retrieves data without modifying state.
Register the MCP server in PolicyLayer and add a rule for ai.extract: 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 Mcp. Nothing to install.
ai.extract 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 ai.extract 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 ai.extract. 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.
ai.extract is provided by the MCP server (@2sio/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 →