AI agents call narrate_audit to retrieve information from Odgs without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool appears designed to retrieve and narrate audit trail information for compliance reporting, which is a read operation. However, confidence is lowered from 0.85 to 0.7 because the description is empty, leaving ambiguity about whether the tool might have side effects (e.g., generating reports that trigger notifications or state changes).
From the tool's definition Tool name 'narrate_audit' suggests narrative generation or reporting on audit data; no verb like 'delete', 'execute', 'pay', or 'modify' indicates destructive, financial, or write operations.
Attacks that exploit this kind of access
narrate_audit. It is categorised as a Read tool in the Odgs MCP Server, which means it retrieves data without modifying state.
Register the Odgs MCP server in PolicyLayer and add a rule for narrate_audit: 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 Odgs. Nothing to install.
narrate_audit 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 narrate_audit 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 narrate_audit. 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.
narrate_audit is provided by the Odgs MCP server (metricprovenance/odgs-mcp-server). 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 →