Dietary Behavior Health Analyzer

Problem Addressed

Everyday eating patterns—eating too fast, poor nutrient balance, inconsistent meal timing—are hard to audit from memory alone. This skill turns video material into a reviewable behavioral observation: it flags tendencies such as wolfing down food, picky eating, distracted eating, or abnormal posture, and also breaks down visible food types, portion sizes, and cooking methods to help generate a structured dietary health reference report.

How It Works

Input can be a local video, image, file path, or web URL. The skill first parses the eating process, then evaluates several dimensions:

• Eating speed: classifies as too fast, moderate, or too slow.
• Meal habits: observes focus, distractions, and posture.
• Food structure: identifies proportions of staples, protein, vegetables, oils, and similar items.
• Risk flags: highlights binge eating, irregular timing, picky eating, or excessive restriction.

History lookup uses --list, reads from the cloud API, and outputs a Markdown table. Analysis depth can be targeted with --analysis-type for comprehensive, speed, habit, structure, or risk views.

Boundaries

The output is a dietary behavior reference, not a diagnosis from a dietitian or physician. Videos should be mp4, avi, or mov and no larger than 10MB. For clear nutritional concerns, long-term restriction, or abnormal eating behavior, seek professional evaluation.