ad-campaign-management
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill performs competitive intelligence by extracting and analyzing ad copy from external libraries (e.g., Facebook Ad Library, LinkedIn Ad Library).
- Ingestion points: Untrusted data enters the agent context through competitor ad copy extracted during research tasks (SKILL.md).
- Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings for the agent when processing external ad content.
- Capability inventory: The skill has file-write capabilities, saving ad drafts to the local content/ directory.
- Sanitization: There is no mention of sanitizing or filtering the content retrieved from external ad platforms before analysis.
- [SAFE]: The skill provides guidance on using well-known AI services (Google Gemini, ElevenLabs, Replicate) and includes code examples that correctly use environment variables ($GEMINI_API_KEY, $ELEVENLABS_API_KEY) for authentication, adhering to secret management best practices.
- [SAFE]: Documented network operations target official and well-known service endpoints for generative media tasks.
Audit Metadata