brida-reflex-duplicate-report-match
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided report summaries and reproduction details, which creates a potential surface for indirect prompt injection. This is mitigated by the skill's specific instruction to only process non-sensitive evidence and the enforcement of a recommendation-only authority boundary that prevents any downstream mutations. Ingestion points: Environment state and report summaries. Boundary markers: Instructions to use bounded evidence. Capability inventory: No file-writing or system-level capabilities. Sanitization: Relies on data normalization.
- [EXTERNAL_DOWNLOADS]: The skill is designed to fetch availability and configuration data from vendor-controlled resources to ensure accurate implementation claims. These downloads are informational and originate from the skill author's infrastructure.
Audit Metadata