AI agents call admin.audit to retrieve information from Signomy without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The name 'admin.audit' suggests a read-oriented auditing or log-review function, consistent with other admin.* tools on this server (reviews, stats, stakes) which appear to be read/review operations. However, the empty description lowers confidence significantly. Severity is medium because admin-level access to audit trails could expose sensitive operational data if misused.
From the tool's definition Tool name 'admin.audit' — 'audit' typically implies reading/reviewing logs or records; description is empty and uninformative.
Attacks that exploit this kind of access
admin.audit. It is categorised as a Read tool in the Signomy MCP Server, which means it retrieves data without modifying state.
Register the Signomy MCP server in PolicyLayer and add a rule for admin.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 Signomy. Nothing to install.
admin.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 admin.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 admin.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.
admin.audit is provided by the Signomy MCP server (sunrisesillneversee/agent-universe). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
admin.audit is one line of Signomy's registry record.
The record carries the whole server: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.
Teams ship this data inside their own products. See what a licence covers →