customer-research
Warn
Audited by Snyk on Jun 14, 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 skill’s REQUIRED workflow includes “Mode 2: Go Find Research” which instructs the agent to gather and extract verbatim quotes from outsider-authored public web content at runtime (e.g., Reddit/G2/app store/forum posts), which then becomes readable LLM context.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata