surveymethods

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.70). The skill connects to SurveyMethods via the Membrane CLI and instructs the agent to list and run actions that read SurveyMethods data (surveys and responses), which are untrusted, user-generated third‑party contents that the agent is expected to interpret and could materially influence subsequent actions.