geo-brand-mentions
Warn
Audited by Snyk on Jul 15, 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). SKILL.md describes runtime WebFetch/Python requests to public sites (e.g., YouTube/Reddit/Wikipedia) and then uses the fetched page/text to assess sentiment/presence, which would feed outsider-authored free text into the agent’s LLM context via those web-scraped results.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata