agentleFS
MuvonMuvon/octomind-tap/skills/chef-budget-meals/SKILL.md

chef-budget-meals

Skill6 starsChanged 8 months agoApache-2.0View on GitHub

Weekly budget meal planner: sale and seasonal price research, grocery list by whole packs, leftovers ledger. Use for a meal plan on a food budget.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask which meals the budget covers.

    line 31
  2. Very specific:

    Use web search for current local retailer prices, promotion dates, pack sizes and regional seasonality.

    line 36
  3. Distinctive:

    Prefer retailer listings or dated flyers for the correct location.

    line 37
  4. Distinctive:

    Never invent prices, specs, laws or references.

    line 39
  5. Distinctive:

    No live price available: ask for a current flyer, receipt or the user's prices.

    line 40
  6. Distinctive:

    Don't claim the total fits while unresolved costs could exceed the cap.

    line 50
  7. Distinctive:

    Before delivering, reconcile servings, ingredient quantities, pack costs, leftover portions, storage windows and budget headroom.

    line 51
  8. Distinctive:

    Never allocate the same portions twice.

    line 56

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Overview

Turns a stated grocery budget and household into a week of affordable meals, a shopping list priced the way the checkout will charge it, and leftovers that are eaten safely. It applies when the user gives a food budget or asks for cheap, affordable or frugal meals. The result is a plan whose total is verified against the cap, or a plainly labelled unpriced draft.

Mental model

Optimize the checkout bill, not a fantasy cost built from spoonfuls of ingredients nobody owns. Saving money is practical, not a moral test: be encouraging, explain the tradeoff (fewer store trips, less waste, a cheaper ingredient doing the same job), and respect familiar foods, limited equipment and the time the person really has. The session's food-safety and allergy rules apply unchanged; savings never override them.

Rules

Intake

  • Establish budget, currency, location, shopping week, household portions, meals covered, allergies, preferences, pantry quantities, stores, transport, equipment, and cooking and storage capacity.
  • Ask which meals the budget covers. Seven dinners are not a whole week's food budget.
  • Scale child portions from the user's guidance, not assumed calorie targets.

Price research

  • Use web search for current local retailer prices, promotion dates, pack sizes and regional seasonality. Read pages with knowledge(command="read", source="<url>"). Cite links and the date checked.
  • Prefer retailer listings or dated flyers for the correct location. Record currency, pack size, unit price, availability, sale expiry, and membership or multibuy conditions.
  • Compare seasonal produce with frozen and canned on usable yield. In season or on sale does not automatically mean cheaper.
  • Never invent prices, specs, laws or references. Separate sourced facts, user-provided figures, assumptions and unknowns.
  • No live price available: ask for a current flyer, receipt or the user's prices. Otherwise deliver an explicitly unpriced draft, not made-up numbers or a guaranteed saving.

Building and pricing the week

  • Choose overlapping ingredients. Schedule perishables first and freeze later portions promptly.
  • Subtract confirmed pantry stock only. Unconfirmed pantry staples are not free.
  • Charge the full pack at checkout even when only part is used, so round purchases up to whole packs. Consumed-ingredient cost may be shown separately.
  • Include known taxes, deposits, delivery charges and required extras. Flag unknown charges.
  • Total the actual basket and reserve a stated contingency inside the budget.
  • Over the cap: swap ingredients or simplify dishes before cutting required servings. If the budget is infeasible, say so and name the constraint to relax.
  • Don't claim the total fits while unresolved costs could exceed the cap. Don't claim verified affordability without a fully priced basket and disclosed conditions.
  • Before delivering, reconcile servings, ingredient quantities, pack costs, leftover portions, storage windows and budget headroom.

Leftovers ledger

  • For each batch track yield, first meal, reserved portions, next use, and the freeze and thaw plan.
  • Never allocate the same portions twice. Never rely on unlimited freezer space.

Output shape

Extend the session's meal-plan shape (week grid, grocery list, prep-ahead) with these parts:

  • Brief: household, covered meals, week and location, currency, spending cap, pantry assumptions, unresolved constraints.
  • Week grid: day, dish, servings, active time, and new, repeat, leftovers or assembly tag.
  • Priced grocery list by store section: item, quantity needed, pack size, packs purchased, price, line total, source and date. Keep pantry use and unpriced items in separate lists.
  • Budget check: basket subtotal, known charges, contingency, final total, headroom, and any per-serving costs with their qualifiers.
  • Prep and leftovers ledger: batch yield, allocated meals and portions, refrigeration deadline, freezing and reheating instructions.
  • Sources and substitutions: linked price, seasonality and safety evidence, sale conditions, cheaper fallbacks.

Save to meals-<slug>.md with text_editor when asked. Buying, paying, posting, sending and deleting stay draft-only until the user says yes to that specific action.

Running without a human

  • Chatting directly: with a clear brief, deliver one best-fit plan, not a menu of options; for budget plans this replaces the session's pick-an-option step. For a material gap ask one focused question, allergies first, then which meals the budget covers.
  • Brief marked AUTONOMOUS or non-interactive: ask nothing. State assumptions, flag allergies as unconfirmed, make no safety claim that depends on them, and leave prices you cannot verify unpriced.

Examples

Bad: "Weekly budget 50, so here are 7 dinners at about 7 each." Good: Ask which meals the 50 covers. Plan breakfasts, lunches and dinners within it, and total the basket.

Bad: "Rice costs 0.20 for the 100 g used, so the list is 14.80." Good: "Rice 1 kg bag, 1 pack at 2.10, charged in full. Basket 49.20 plus 2.00 contingency, headroom 0.80. Checked 4 Nov, flyer link below."

Bad: "Chili makes 8 portions: Monday dinner, Tuesday lunch, Tuesday dinner." Good: Ledger: 8 portions. Mon dinner 2, Tue lunch 2, freeze 4 (next use Fri, thaw in fridge Thu). Total 8.

Checklist

  • Budget scope (which meals) and allergies confirmed, or flagged as assumed?
  • Every price has a source link and date, or is marked unpriced?
  • List priced by whole packs, with confirmed pantry stock subtracted?
  • Taxes, deposits and delivery included or flagged unknown?
  • Contingency and headroom stated, final total under the cap?
  • No portion appears twice in the leftovers ledger?
  • Servings, pack costs and storage windows reconciled?
  • Output includes brief, grid, priced list, budget check, ledger, sources?

Composition / References

  • Storage windows and temperatures: cite current USDA FSIS or UK FSA guidance for the user's country.
  • The session's food-safety, allergy and substitution rules apply unchanged.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.