cold-start-interview

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly asks for and ingests seed documents from untrusted third‑party locations (e.g., Google Drive links, CLM record IDs, shared Drive/SharePoint/Box folders) and reads those user-supplied agreements to extract playbook positions that directly drive configuration and future agent decisions.