ai-marketing
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes a implementation template for sentiment analysis that processes large quantities of external, untrusted customer comments.
- Ingestion points: SKILL.md (Step 5, Use Case 1: Sentiment Analysis at Scale) where user comments are interpolated into a prompt.
- Boundary markers: The provided prompt template uses a '[Comment data]' placeholder but lacks explicit instructions for the AI to disregard any command patterns embedded within that data.
- Capability inventory: The skill defines frameworks for data analysis, trend prediction, and automated content generation; no executable scripts or code files are included in the skill.
- Sanitization: The instructions do not mention filtering, escaping, or sanitizing external comments to mitigate the risk of hidden prompt injection payloads.
Audit Metadata