keyword-research
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.80). The skill's research methods explicitly direct mining and interpreting untrusted public sources—e.g., Google Autocomplete/People Also Ask and site:reddit.com and site:quora.com searches—so the agent would read user-generated third-party content that could carry indirect prompt injections.
Audit Metadata