customer-research

Warn

Audited by Snyk on Aug 22, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (medium risk: 0.30). In required runtime workflow “customer-research,” Mode 2 (“Go Find Research” / “Digital Watering Hole Research”) directs the agent to search and read outsider-authored online community/review content (e.g., Reddit, G2/Capterra, app reviews) and extract verbatim quotes from whatever it finds.

Issues (1)

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 22, 2026, 11:35 AM
Issues
1