TCM Facial Diagnosis Analysis

Problem

In video analysis workflows, a user may have only a facial video but wants a structured TCM-style assessment: constitution, complexion, organ tendencies, health flags, and wellness advice. This skill turns a video URL or local MP4/AVI/MOV file into a readable report and queries historical reports from a unified cloud source.

How It Works

• Video diagnosis: upload a local video or provide a public URL; scripts/face_analysis.py calls the API. Local files are sent via multipart/form-data; public URLs are downloaded by the service.
• Structured output: the report typically includes overall constitution, organ condition, complexion analysis, health warnings, and wellness advice. --detail can produce basic, standard, or json output.
• History listing: when the user asks for historical TCM reports, report lists, or report tables, use --list --open-id to query the cloud API and render a Markdown table with Report Name, Analysis Time, and View.
• Key constraints: a valid open-id is required before analysis and must not be guessed or generated; historical data must come from the cloud API, not local memory, MEMORY.md, or LanceDB long-term storage.

Boundaries

This skill is scoped to TCM facial diagnosis and does not perform fall detection, behavior recognition, or general risk analysis. Results are for reference only and do not replace professional medical diagnosis. Videos should be mp4/avi/mov, up to 100MB, and the skill’s bundled scripts should be used instead of ad-hoc replacements.