testing-with-sandbox

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs reading messages from the Email Sandbox API (e.g., "List messages" and "Fetch a message" endpoints like GET https://mailtrap.io/api/accounts/{account_id}/inboxes/{inbox_id}/messages/{id}) and states you should "run integration or CI checks that read from the Email Sandbox API," meaning the agent will ingest untrusted, user-generated email content that could influence its actions.