social-listening-brief
Warn
Audited by Snyk on Jul 2, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.95). The skill’s runtime workflow explicitly ingests outsider-authored public free text via UnifAPI endpoints that return verbatim posts/comments/articles (e.g.,
x/tweets/search/recent,reddit/posts/{id}/comments,youtube/searchtitles/descriptions/captions,tiktok/searchcaptions,threads/search/*,news/search), which the agent then clusters and quotes into the LLM context.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata