customer-research

Warn

Audited by Snyk on Apr 23, 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's "Review Mining" section explicitly directs the agent to extract insights from public, user-generated sources (G2, Capterra, Reddit, Twitter/social mentions, app store reviews and support forums) and the Mandatory Output Rules require citing specific sources, meaning the agent is expected to ingest and act on untrusted third-party content that can influence decisions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 23, 2026, 12:49 PM
Issues
1