generated-media-qa
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external media QA reports through the
scripts/normalize_qa_report.pyscript. This represents an ingestion point for untrusted data that could potentially influence agent behavior if malicious instructions are embedded in the report fields. - Ingestion points: The
scripts/normalize_qa_report.pyscript reads JSON data from local files provided via command-line arguments (e.g.,report.json). - Boundary markers: The skill instructions recommend separating documented facts, empirical observations, and heuristics. The normalization script enforces a specific JSON schema and distinguishes evidence types.
- Capability inventory: The script is capable of reading and writing local JSON files for normalization. It does not perform network operations, access sensitive system paths, or execute arbitrary system commands.
- Sanitization: The script implements validation for field types, date formats (ISO-8601), and specific enumerated values for severities, evidence types, and dispositions.
Audit Metadata