zendesk

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

• Third-party content exposure detected (high risk: 0.70). This skill calls Zendesk endpoints such as /api/v2/tickets.json and /api/v2/tickets/{id}/comments.json and ingests ticket subjects/comments and user messages (user-generated content) which the agent is expected to read and could contain untrusted/malicious instructions.