create-verification-skill
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a structured approach for automating user-level application verification. It includes patterns for isolated test environments, diagnostic 'doctor' checks, and evidence collection as demonstrated in the references directory.\n- [INDIRECT_PROMPT_INJECTION]: The skill instructions create a surface for indirect prompt injection by design, as the agent must ingest repository content to generate verification logic. While this is the intended primary purpose, the following assessment is noted:\n
- Ingestion points: Project routes, CLI commands, and documentation discovered within the repository (SKILL.md, Step 1).\n
- Boundary markers: None specified for the generated skill content.\n
- Capability inventory: The generated skill is intended to execute application startup and test commands (SKILL.md, Step 2).\n
- Sanitization: None specified for repo-derived commands.
Audit Metadata