recipes

name	recipes
description	Generate meal plans and recipes from optimization results. Use when user wants recipes, meal ideas, cooking instructions, or a weekly meal plan from their optimized food list. Triggers on "create recipes", "meal plan recipes", "how to cook", "weekly menu", "what to make", "cooking ideas".

Recipe Generation from Optimized Foods

This skill takes optimization results and creates practical meal plans with recipes. It can either generate a prompt for external use or create recipes directly.

Best Source: Template-Based Optimization

Template-based results are ideal for recipes because they're already structured as proper meals:

# Run template-based optimization first
uv run llmn optimize --pattern pescatarian --pattern slow_carb --template --json

This gives you:

• Breakfast: Eggs 150g, Edamame 100g, Kale 100g
• Lunch: Cod 235g, Kidney beans 216g, Zucchini 161g
• Dinner: Salmon 170g, Lentils 113g, Carrots 149g
• Snack: Peanuts 30g

No meal allocation needed - foods are already assigned to meals with proper structure.

Workflow

Step 1: Get the Optimization Results

Check if there's a recent optimization run:

# List recent optimization runs
uv run llmn explore runs --limit 5 --json

If no recent runs, suggest running /llmn or /multiperiod first.

To use a specific run:

# Get details of a specific run
uv run llmn explore runs --json
# Look at run_id in the output

Step 2: Export the Food List

Export the optimization results in a format suitable for recipe generation:

# Export latest run for LLM recipe generation
uv run llmn export-for-llm latest

# Or export with specific number of days
uv run llmn export-for-llm latest --days 7

# Export to a file
uv run llmn export-for-llm latest --days 7 --output /tmp/meal_prompt.md

Step 3: Gather User Preferences

Ask the user about their cooking preferences:

1. Cuisine styles (select multiple):

   • Mediterranean
   • Asian
   • Mexican/Latin
   • American
   • Indian
   • Italian
   • Middle Eastern

2. Cooking skill level:

   • Beginner (simple recipes, <20 min)
   • Intermediate (moderate complexity, <30 min)
   • Advanced (willing to try complex techniques)

3. Max prep time per meal:

   • Quick (<15 min)
   • Moderate (<30 min)
   • Extended (30-60 min)

4. Available equipment:

   • Stovetop
   • Oven
   • Instant Pot / Pressure Cooker
   • Air Fryer
   • Slow Cooker
   • Blender
   • Grill

5. Meal prep preference:

   • Cook fresh daily
   • Batch cook on weekends
   • Mix of both

Step 4: Generate Recipes

Option A: Generate a prompt (user will paste elsewhere)

uv run llmn export-for-llm latest --days 7

Then combine with user preferences to create a detailed prompt.

Option B: Generate recipes directly (Claude creates recipes now)

Use the optimization output to create recipes that:

• Use EXACTLY the specified food amounts
• Distribute foods across meals appropriately
• Match user's cuisine preferences
• Fit within time/skill constraints
• Consider meal prep strategies

Step 5: Present the Meal Plan

Structure the output as:

## Day 1

### Breakfast
**Scrambled Eggs with Spinach** (10 min)
- 150g eggs (3 large)
- 50g spinach
- 15g olive oil

Instructions:
1. Heat olive oil in a pan
2. Add spinach, cook until wilted
3. Add beaten eggs, scramble until set

Macros: 350 kcal, 25g protein, 2g carbs, 26g fat

---

### Lunch
**Salmon with Lentils** (25 min)
...

### Dinner
**Tuna and Black Bean Bowl** (20 min)
...

### Snack
**Almonds** (0 min)
- 25g raw almonds

Macros: 145 kcal, 5g protein, 5g carbs, 12g fat

---

## Daily Totals
- Calories: 1850 kcal
- Protein: 155g
- Carbs: 120g
- Fat: 78g

Recipe Generation Guidelines

Respect Food Amounts

The optimization calculated exact amounts. Recipes should use these amounts:

• If optimization says 150g salmon, recipe uses 150g salmon
• Scale across multiple days if doing weekly meal prep
• For 7-day plan: daily amount × 7 = weekly shopping amount

Food Preparation Context

The export includes preparation state hints:

• Salmon (raw) → needs cooking
• Lentils (cooked) → already cooked, can use canned
• Eggs (whole, raw) → needs cooking
• Spinach (raw) → can be eaten raw or cooked

Meal Allocation Logic

Skip this if using template-based optimization - meals are already structured.

If using single-period optimization (not template or multi-period), allocate foods to meals sensibly:

Breakfast foods:

• Eggs
• Oatmeal
• Yogurt
• Fruit

Lunch/Dinner foods:

• Fish (salmon, tuna, cod)
• Legumes (lentils, beans, chickpeas)
• Vegetables (broccoli, spinach, cauliflower)
• Grains (rice, quinoa)

Snack foods:

• Nuts (almonds, walnuts)
• Small portions of fruit
• Vegetables with hummus

Recipe Complexity by Skill Level

Beginner:

• One-pan meals
• No-cook options (salads, bowls)
• Simple techniques: scramble, boil, roast
• Example: "Scrambled eggs", "Baked salmon with roasted vegetables"

Intermediate:

• Multi-component meals
• Sauces and marinades
• Example: "Pan-seared salmon with lemon-herb sauce and garlic lentils"

Advanced:

• Complex techniques
• Multiple cooking methods
• Example: "Sous vide salmon with crispy skin, served with French lentils du Puy"

Meal Prep Strategies

For batch cooking:

• Proteins: Cook all proteins Sunday, refrigerate
• Grains/Legumes: Make large batch, portion into containers
• Vegetables: Prep (wash, chop) Sunday, cook fresh or roast in bulk
• Sauces: Make dressings/sauces to add variety to same base foods

Example meal prep plan:

SUNDAY PREP (2 hours):
- Bake 1kg salmon (5 portions)
- Cook 500g lentils (5 portions)
- Roast 1kg mixed vegetables
- Hard-boil 12 eggs
- Wash and portion spinach

WEEKDAY ASSEMBLY (5-10 min):
- Combine pre-cooked components
- Add fresh elements (avocado, olive oil)
- Microwave to reheat if needed

Commands Reference

# List recent optimization runs
uv run llmn explore runs --limit 10 --json

# Export latest run for recipes
uv run llmn export-for-llm latest

# Export specific run
uv run llmn export-for-llm <run_id>

# Export with day count
uv run llmn export-for-llm latest --days 7

# Export to file
uv run llmn export-for-llm latest --output /tmp/recipes.md

# Get optimization details (for macro breakdown)
uv run llmn explore runs --json

Example Output

Prompt Generation Mode

# Meal Planning Request

## Optimized Daily Food Allocation

| Food | Amount | Protein | Calories |
|------|--------|---------|----------|
| Salmon, Atlantic, raw | 150g | 31g | 310 |
| Eggs, whole, raw | 150g | 19g | 215 |
| Lentils, cooked | 200g | 18g | 232 |
| Spinach, raw | 100g | 3g | 23 |
| Broccoli, raw | 150g | 4g | 51 |
| Olive oil | 30g | 0g | 265 |
| Almonds | 25g | 5g | 145 |

**Daily Totals**: 1850 kcal, 155g protein

## Preferences
- Cuisine: Mediterranean, Asian
- Skill level: Intermediate
- Max prep time: 30 minutes
- Equipment: Stovetop, oven, air fryer
- Meal prep: Batch cook on weekends

## Request
Create a 7-day meal plan using exactly these daily food amounts.
Distribute across breakfast, lunch, dinner, and one snack.
Include simple recipes that reheat well.

Direct Recipe Mode

Generate recipes directly based on the food list and preferences, structured as shown in Step 5.