Infant Prone Sleeping Asphyxia Alert

Problem

Infant sleep monitoring can easily miss short-lived but dangerous states: prone sleeping, head covering, or bedding and clothing blocking the nose and mouth. This skill takes an infant sleep monitoring video and turns visual posture and facial coverage signals into a readable risk report, reducing the chance that a parent misses a high-risk moment.

How It Works

• Input: accepts a local video file or a public video URL, commonly mp4, avi, or mov, with a file size around 10MB.
• Detection: identifies high-risk factors such as prone sleeping, head covered by bedding, and nose or mouth occlusion.
• Alerting: when a high-risk state is detected, it produces a structured report covering posture, head covering, nose or mouth occlusion, risk level, and alert message.
• History lookup: uses --list to fetch past warnings from the cloud API and render them as a Markdown table with report name, analysis time, risk level, and a view link.

Limits

This is an auxiliary monitoring tool and does not replace adult supervision. open-id is required to save and query the current user's warning records. Historical data is fetched from the cloud interface rather than local memory. If the video angle is poor, the face is blocked, or the scene is ambiguous, the result may be limited. If a high-risk state is reported, check the infant immediately and seek medical help if needed.