Recipe Validation Skill

Purpose

Validate recipe data against the schema and quality standards before committing changes.

Activation Triggers

•Modifying data/recipes.json
•Keywords: validate, check, schema, verify

Validation Command

bash

python scripts/validate-recipes.py

For strict mode (fails on warnings):

bash

python scripts/validate-recipes.py --strict

What Gets Validated

Required Fields

•id - Unique slug identifier
•collection - Must be "all" for this repo
•title - Recipe name
•category - Must be from valid list
•ingredients - Array with at least one item
•instructions - Array with at least one step

Optional Fields (Validated if Present)

•source_note - Source attribution
•description - Brief description
•servings_yield - Number of servings
•prep_time, cook_time, total_time
•temperature - Cooking temperature
•pan_size - Required pan/dish
•notes - Array of strings
•tags - Array of tag strings
•confidence - Object with overall and flags
•image_refs - Array of filenames
•nutrition - Nutrition data object

Valid Categories

code

appetizers | beverages | breads | breakfast | desserts
mains | salads | sides | soups | snacks

Sanity Limits

The validator checks for implausible values:

Check	Limit	Why
Salt	≤ 4 tbsp	More is likely an error
Sugar	≤ 8 cups	Reasonable max for large batches
Temperature	200-550°F	Normal cooking range
Servings	1-100	Reasonable yield range

Error Types

Errors (Must Fix)

•Missing required field
•Invalid category
•Duplicate recipe ID
•Malformed JSON
•Invalid field type

Warnings (Review)

•Missing optional fields (source_note, confidence)
•Sanity limit exceeded
•Unusual ingredient quantities
•Missing image_refs

Resolution Workflow

Step 1: Run Validation

bash

python scripts/validate-recipes.py

Step 2: Review Output

code

Validating recipes.json...
  ✓ Recipe "chocolate-cake" valid
  ✗ Recipe "mystery-dish" - Missing required field: category
  ⚠ Recipe "salty-soup" - Warning: salt quantity (5 tbsp) exceeds limit

Step 3: Fix Errors

•Missing fields: Add the field with appropriate value
•Invalid category: Use category from valid list
•Duplicate ID: Make IDs unique
•Sanity warnings: Verify against source image

Step 4: Re-validate

bash

python scripts/validate-recipes.py

Step 5: Commit Only When Clean

Only commit when validation passes without errors.

Guardrails

MUST

•Run validation before every commit
•Fix all errors before proceeding
•Investigate warnings (don't ignore)
•Verify fixes against source images

MUST NOT

•Commit with validation errors
•Ignore sanity limit warnings without verification
•Invent data to pass validation
•Disable or bypass validation

Integration with Hooks

The post-write-validate.sh hook automatically runs validation when recipes.json is modified. Check the output for any errors.

Accuracy is more important than speed.