analytics-insights
Warn
Audited by Snyk on Feb 27, 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). The SKILL.md workflow explicitly instructs the agent to consult and interpret open/public third-party sources (e.g., "check SEMrush Sensor, MozCast, Google Search Console," "Meta Ad Library," "Reddit/Quora," "Wayback Machine," competitor emails and review sites) as part of anomaly diagnosis and competitive intelligence, exposing the agent to untrusted user-generated content that can influence decisions and next actions.
Audit Metadata