AI agents call extract_by_selector to retrieve information from Crawler without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool reads and parses already-fetched web page content to extract information via CSS selectors. It performs no write operations, does not execute code or scripts, and cannot modify or delete data. The only risk is potential information disclosure if sensitive data is inadvertently exposed on a page, but this is a consequence of the page content itself, not the tool's design.
From the tool's definition Tool description states 'Extract specific data from a page using a CSS selector.' This is a pure data retrieval operation with no modification, deletion, or execution capabilities.
Attacks that exploit this kind of access
Extract specific data from a page using a CSS selector. It is categorised as a Read tool in the Crawler MCP Server, which means it retrieves data without modifying state.
Register the Crawler MCP server in PolicyLayer and add a rule for extract_by_selector: 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 Crawler. Nothing to install.
extract_by_selector 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 extract_by_selector 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 extract_by_selector. 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.
extract_by_selector is provided by the Crawler MCP server (shadab15github/crawler-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 →