review-harness
Warn
Audited by Snyk on Aug 16, 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 skill’s runtime review workflow reads a user-supplied “replay archive” zip and inspects free-text model outputs in
steps[*][*].action.thoughtsplusactionString, so an outsider can poison by submitting arbitrary replay JSON text that the reviewer LLM processes.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata