separate-verification-validation
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script
scripts/validate-vv-evidence.pyto analyze and validate markdown evidence files produced by the agent. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, including requirement lists and user scenarios, to populate evidence tables. While the validation script checks for structural completeness, this data ingestion represents a potential attack surface for indirect prompt injection from the documentation being reviewed.
Audit Metadata