AI agents call get_audit_report to retrieve information from Web Audit without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and returns previously-generated audit reports. No data is created, modified, deleted, or executed. The blast radius is minimal — an agent can only access audit data that has already been computed and stored, which poses no risk of unintended side effects or resource consumption.
From the tool's definition Tool name is 'get_audit_report' and description states 'Fetch a persisted audit report' — fetch is a read operation that retrieves existing data without modification or side effects.
Attacks that exploit this kind of access
Fetch a persisted audit report with findings and metrics. It is categorised as a Read tool in the Web Audit MCP Server, which means it retrieves data without modifying state.
Register the Web Audit MCP server in PolicyLayer and add a rule for get_audit_report: 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 Web Audit. Nothing to install.
get_audit_report 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 get_audit_report 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 get_audit_report. 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.
get_audit_report is provided by the Web Audit MCP server (sayuru-akash/web-audit-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 →