kami-suspicious-person

• Analyzed: May 13, 2026, 7:51 AM
• Findings: 1
• Version: 1.0.1
• Hash: 6a9813e904314f18d6d887fc7ebb13d01c29b498e043ea52d58c719b0d3a403f
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill implements a legitimate face-recognition-based loitering detection system using ONNX models (SCRFD and ArcFace) for local CPU inference. It includes a setup script that fetches models from the official InsightFace GitHub repository and a main detection script that monitors RTSP streams, tracks strangers, and sends alerts via Feishu webhooks and a local 'inbox' file. While the SKILL.md contains 'Strict Rules' that use authoritative language to direct the AI agent's behavior (e.g., requiring background execution and specific heartbeat reporting logic), these instructions are clearly aligned with the functional requirements of a continuous monitoring service and do not exhibit malicious intent or hidden data exfiltration.