Plan household meals with AI
Keep declared food preferences, recipes and shopping plans together while recording cooked and eaten meals as separate actual events.
Plan household meals with AI without repeating every declared preference at the start of a conversation. The meal-magic skill uses household context to suggest food and can retain recipes, multi-day plans and shopping lists. Ingredients mentioned for tonight do not automatically become permanent pantry stock or long-term preferences.
Suggestions, saved plans and actual events stay separate. Cooking or eating enters history only after an explicit report, and a proposed diner does not prove participation. Data is stored locally, but the hosting agent may process what it reads through its own model provider. Agents sharing the same operating-system user and data directory share one profile, rather than isolated encrypted accounts.
What it does
- Use declared household preferences in suggestions
- Prepare shopping lists for selected plans
- Retrieve saved recipes without silently rewriting them
- Record cooked and eaten events separately
- Review meal history for variety
- Correct or forget a stated preference
Install and use
- Install the skill for a file-capable agent with Python3.10 or newer. Choose an authorised data location and review the host model's privacy settings; no dedicated cooking API key is needed.
- Describe the request: “Suggest dinner from these ingredients without saving them as permanent preferences.”
- Explicitly request recipe or plan saving and check that it succeeds.
- Report actual cooking or eating separately, naming any record you want corrected or removed.
FAQ
Source
Made by alexneverland. Original skill: meal-magic, published on GitHub under the MIT licence.
View on GitHub