research-general
Warn
Audited by Snyk on May 12, 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 SKILL.md explicitly instructs using WebSearch/WebFetch to collect and read content from public websites (and community sources like Stack Overflow and Reddit shown in references/source-evaluation.md), which the agent must interpret and use to form conclusions and reports.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata