convex-cost
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external telemetry data from Convex deployment insights and table structures to perform its analysis, creating an attack surface for indirect prompt injection.
- Ingestion points: External data is collected from the
insights(bytes/documents-read) andtables(row counts) tools during the GATHER workflow step. - Boundary markers: The instructions do not define specific delimiters or "ignore previous instructions" markers when interpolating telemetry data into the analysis context.
- Capability inventory: The skill's actions are restricted to analysis and emitting findings to an internal bus; it does not possess capabilities for arbitrary command execution, file system writes, or unauthorized network operations.
- Sanitization: There is no mention of sanitization or validation of the ingested telemetry strings before they are processed by the agent.
Audit Metadata