videodb

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill contains explicit, high-risk surveillance and exfiltration capabilities (desktop screen + mic capture, real-time visual/text indexing including detecting password fields, and alerts/webhook callbacks that can send data to arbitrary external endpoints) which can be used as a covert backdoor to steal sensitive information even though the code itself has no obfuscated payloads execs, or explicit reverse-shells.

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests arbitrary public URLs/YouTube videos, RTSP/RTMP streams, and live desktop captures (see "Common inputs" in SKILL.md and upload/connect_rtstream examples in reference/rtstream-reference.md and capture.md), then indexes/transcribes and runs AI analyses on that untrusted content to drive searches, timeline edits, alerts, and webhooks — meaning third-party content can directly influence agent actions.