data-monitoring
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from local product documentation (
docs/product/prd/*.md) and external search results to define metrics and benchmarks. While it lacks explicit boundary markers or sanitization for this content, its primary capabilities are limited to file reading and structured data writing (docs/product/.ompm/metrics.json), which significantly limits the impact of potential injection attacks. - Ingestion points:
docs/product/prd/*.md, Web search results (for industry benchmarks). - Boundary markers: Absent.
- Capability inventory: File reading, JSON file writing.
- Sanitization: Absent, though the skill mandates source verification and confidence ratings for all external data.
Audit Metadata