customer-research
Warn
Audited by Snyk on Aug 19, 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). The “customer-research” skill’s runtime workflow supports Mode 2 “Go Find Research” by reading/verbatim-extracting outsider-authored content from online sources like Reddit and G2 (via the included “Digital Watering Hole Research” and per-platform source guides), meaning an outsider can post free text into those feeds/queues for the agent to ingest.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata