convex_audit_schema

What convex_audit_schema does on Nodebench

AI agents call convex_audit_schema to retrieve information from Nodebench without modifying anything. It is typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.

Why convex_audit_schema is rated Low

Analysis tools retrieve and examine data structures without modifying them, fitting the Read category. Severity is medium rather than low because schema analysis could expose sensitive information about data models, authentication patterns, or business logic embedded in a Convex backend, which could inform further attacks. However, it performs no direct data access, code execution, or destructive action on its own.

From the tool's definition Tool name and description indicate 'Analyze a Convex project' — an inspection/analysis operation with no modification capability. Convex is a backend platform, and analyzing a schema is a read-only introspection task.

The rule that runs convex_audit_schema safely

PolicyLayer is an MCP gateway: it sits between your AI agents and Nodebench, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For convex_audit_schema, this is the rule to start with:

{
  "version": "1",
  "default": "deny",
  "tools": {
    "convex_audit_schema": {}
  }
}

Questions about convex_audit_schema