Plan everyday meals with AI
Use confirmed information to choose portions and alternatives, keep planned meals separate from actual intake and optionally retain records in a local folder.
Plan everyday meals with AI around the ingredients, time and constraints you actually have. The nutrition-skill package uses a local core to propose portions and check defined ranges from confirmed inputs. It should not infer health facts from a name, photograph or missing answer.
For example, keep an already chosen breakfast fixed while asking for a low-effort lunch from available ingredients. A supported replacement can preserve locked or already eaten meals. The numerical path has defined general-adult eligibility and a limited food catalogue; matching energy and protein ranges is not a complete nutritional, allergy or medical assessment.
What it does
- Suggest meals from confirmed ingredients and constraints
- Swap a meal while preserving fixed choices
- Record actual intake separately from the plan
- Aggregate quantities for a shopping list
- Review progress from genuine recorded information
- Keep local records when explicitly chosen
Install and use
- Provide the complete package to an authorised file-and-command-capable agent with Python3.11 or newer installed. The local core needs no API key; host access is separate.
- State the goal and relevant constraints. Choose session-only use or an authorised data folder outside the code.
- Review proposed options and quantities, keeping unknown information unresolved rather than guessing. Identify meals that must stay fixed.
- Record what was actually eaten and review later changes without overwriting that history.
FAQ
Source
Made by aiiqc / Nutrition Skill contributors. Original skill: nutrition-skill, published on GitHub under the MIT licence.
View on GitHub