market-research
Warn
Audited by Snyk on Mar 6, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to fetch and interpret open/public, user-generated sources (see Step 3: "Reddit", "G2, Capterra, Trustpilot", "Twitter/X", competitors' websites, etc., and Step 6: "pull from forum posts or interviews"), and to synthesize those findings into decisions and next actions, which meets all criteria for exposure to untrusted third-party content that could enable indirect prompt injection.
Audit Metadata