customer-research
Warn
Audited by Snyk on May 19, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.90). This skill explicitly instructs the agent (see Part 6 "Review Mining" and Part 7 "Digital Watering Holes Research") to fetch and analyze public, user-generated content from G2, Capterra, Reddit, app stores, Twitter, LinkedIn, and similar sources and then use that language to drive messaging and decisions, which exposes the agent to untrusted third-party content that can influence actions.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata