Stroke Risk Screening Analysis

Problem

Stroke risk can appear through subtle facial and constitutional signals that a single blood-pressure reading or routine lab value may miss. This skill combines facial images or video with optional physiological metrics to produce screening signals for high-risk states such as cerebral infarction or hemorrhage, suitable for generating structured risk reports in an agentic workflow.

Core capabilities and workflow

• Facial feature analysis: evaluates complexion, tongue appearance, and mouth posture described as TCM face and tongue signals.
• Risk grading: uses optional inputs such as blood pressure, fasting blood glucose, and total cholesterol to issue graded warnings.
• Advice generation: provides dietary and lifestyle guidance, plus a report link.
• Historical report lookup: uses --list to query cloud-stored reports and render them as a Markdown table.

The typical flow is to provide a local path or --url, set --media-type, and run the skill script. User identity is handled internally and should not be requested from or exposed to the end user.

Boundaries

The output is a health-risk screening prompt, not a substitute for professional medical examination or diagnosis. Inputs should be clear MP4/AVI/MOV video or JPG/PNG/JPEG image files up to 10 MB; if a high-risk signal is reported, seek medical care promptly.