extract-secrets

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow creates a headless RE session for a user-supplied URL (skill_session_create with args.url) and then runs skill_extract_secrets, which ingests and scans the target site's decompiled JavaScript (outsider-authored if the user points it at an attacker-controlled or third-party site).