Scan an LLM response for sensitive data it may have generated, inferred, or echoed.
AI agents call scan_response to retrieve information from Query Sanitizer without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs security auditing by detecting sensitive information in LLM outputs, but does not modify data, execute commands, delete anything, or move money. It is purely observational and diagnostic in nature, similar to a search or query operation. The blast radius of misuse is minimal—worst case it fails to detect sensitive data in a response, but the tool itself cannot cause harm through its execution.
From the tool's definition Tool name is 'scan_response' and description states it 'scan[s]' responses 'for' sensitive data—a read-only inspection operation. The verb 'scan' indicates passive analysis without modifying the response or taking action.
Attacks that exploit this kind of access
Scan an LLM response for sensitive data it may have generated, inferred, or echoed. It is categorised as a Read tool in the Query Sanitizer MCP Server, which means it retrieves data without modifying state.
Register the Query Sanitizer MCP server in PolicyLayer and add a rule for scan_response: 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 Query Sanitizer. Nothing to install.
scan_response 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 scan_response 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 scan_response. 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.
scan_response is provided by the Query Sanitizer MCP server (vidoluco/query-sanitizer-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 →