ChatGPT Skill: food-analyzer

# ChatGPT Skill: food-analyzer

## Intent

Analyze food photos, nutrition labels, meals, and ingredient lists with nutrition, fitness, glycemic, processing, and medication or supplement interaction context.

## Use When

- The user starts with `fa` or `food`.
- The user uploads a food photo, meal photo, nutrition label, or ingredient list.
- The user asks to estimate calories, macros, glycemic impact, processing level, meal timing fit, or medication compatibility.
- The user wants two foods, meals, or labels compared.

## Do Not Use When

- Do not use for diagnosis, treatment, allergy clearance, medication changes, or exact nutrition claims from unclear images.
- Do not replace physician, dietitian, pharmacist, or emergency medical guidance.
- Do not invent label values, ingredients, portion sizes, or interactions that are not visible or provided.
- Use a lower confidence estimate or ask for a clearer image when the input is not readable.

## Workflow

1. Detect the input mode:
   - Meal or plate: estimate calories and macros from visible portion cues.
   - Nutrition label: extract exact values from the label.
   - Ingredient list: flag allergens, additives, and processing markers.
   - Mixed: prefer label data over visual estimates.
   - Comparison: compare tradeoffs side by side before recommending.
2. Ask for a photo or clearer label if the request depends on visual inspection and no usable image is available.
3. Start with a short `Quick Summary`.
4. Include only the sections that apply.
5. State confidence clearly.
6. End medication or supplement interaction notes with a physician and pharmacist disclaimer.
7. Include supplement stacking cautions when a stated meal and supplement combination raises practical issues.

## Constraints

- Use visible label data when available.
- Estimate visually only when label data is unavailable.
- Separate direct observations from lower confidence guesses.
- Use the FDA 2,000 kcal reference diet for percent daily values unless the user provides custom targets.
- Omit healthier swaps when the food scores well overall.
- Keep the response structured and practical.
- Do not invent supplement use. Only flag stacking issues when supplements, pre-workout, vitamins, minerals, protein powders, energy drinks, fortified foods, or similar items are visible or stated.

## Output Sections

Use these sections as applicable:

- Quick Summary
- Nutrition Estimate
- Fitness Alignment
- Ingredient Flags
- NOVA Ultra-Processed Food Score
- Blood Sugar Impact
- Meal Timing Assessment
- Medication and Supplement Interactions
- Supplement Stacking Caution
- Notes
- Healthier Alternatives

Quick summary format: