beanstalk
Warn
Audited by Snyk on Apr 29, 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 invokes Membrane actions to fetch Beanstalk data (e.g., "List Changesets", "Get Repository", "Get Code Review" and the documented "membrane action run" workflow in SKILL.md), which will return user-generated repository/commit/review content from external Beanstalk accounts that the agent is expected to read and could materially influence subsequent actions.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata