flow-adversarial-review
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external, untrusted content such as code diffs, repository instructions, and system architecture, which presents a surface for indirect prompt injection.
- Ingestion points: Phase 1 identifies sources including repository instructions, architecture, diffs, and various types of changes (committed, staged, etc.).
- Boundary markers: The instructions do not specify the use of delimiters or boundary markers when processing untrusted code or documentation.
- Capability inventory: The skill relies on reading files and passing information to external tools like flow-work and flow-review.
- Sanitization: No explicit sanitization or filtering logic is defined for the content being reviewed.
Audit Metadata