audience-behavior-study
Warn
Audited by Snyk on Aug 22, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (medium risk: 0.30). In Step 2 “Research”, the required workflow explicitly uses
WebSearchandWebFetchto pull and quote public threads/reviews from outsider-authored sources (Reddit/X/review sites/niche communities), so injected free text from those posts can be ingested without selecting a specific item first.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata