customer-health-scorecard
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external customer files located in directories like customers/ and accounts/. This creates a vulnerability surface where untrustworthy content within these files could potentially influence the agent's behavior during data processing.
- Ingestion points: Reads from paths like customers/, health/, accounts/, and user-specified scorecard files.
- Boundary markers: No specific delimiters or instructions are used to isolate ingested data from the prompt logic.
- Capability inventory: Performs file reads to gather metrics and file writes to save markdown scorecards.
- Sanitization: No evidence of sanitization or validation of the ingested content prior to interpolation.
Audit Metadata