audience-research
Warn
Audited by Snyk on Jul 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.85). The required workflow explicitly instructs mining “reviews, social comments, Reddit/forums/communities, support tickets/FAQs, sales-call notes, DMs, and search queries” from public sources at runtime, which are outsider-authored free text that would be ingested into the agent’s LLM context for summarization/clustering.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata