customer-research
Warn
Audited by Snyk on Mar 28, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 1.00). The skill's Mode 2 workflow and its references/source-guides.md explicitly instruct the agent to gather and read user-generated content from open public sites (e.g., Reddit, G2/Capterra, Hacker News, YouTube comments, LinkedIn, app stores), meaning the agent will fetch and interpret untrusted third-party content that can influence research conclusions and subsequent actions.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata