Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Cost a menu item to a plate cost and food-cost percentage, then price it for a target margin. Use when asked to cost a dish, calculate food cost percentage, price a menu item, or engineer a menu for profitability. Produces a plate-cost breakdown (ingredient × yield × price), the food-cost %, a suggested price for the target margin, and menu-engineering flags (star / plow-horse / puzzle / dog) so the operator knows what to promote, reprice, or cut.
.claude/skills/mohitagw15856-menu-cost-engineer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 51% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 78% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 86% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 55% | 0% |
Restaurants don't fail on bad food — they fail on math they never did. A dish that "feels" priced right can quietly run a 45% food cost. This skill costs the plate honestly (accounting for yield loss and waste), sets a price to the margin the operator actually needs, and flags where the menu is leaking money.
Given a recipe and ingredient prices, produce the full costing — estimate yields and waste where not given (label the assumption). If prices are missing, ask or use a stated regional estimate and flag it.
Ask for (if not provided, else infer and label):
| Ingredient | As-purchased cost/unit | Yield % | Portion | Plate cost | |---|---|---|---|---| | | | | | |
Include a line for waste/trim and any garnish/consumables.
Given popularity (if provided) and margin, classify: Star (high both — feature it), Plow-horse (popular, low margin — reprice/re-engineer), Puzzle (high margin, low sales — promote/reposition), Dog (low both — cut or rework).
2–3 concrete moves — portion, spec, supplier, price, or plate composition — to hit the target without hurting the guest experience.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 8,011 | 9,668 | +21% | 1 | 1 | 0% | 1,587 | 2,396 | +51% | 0 | 0 | — |
case-02 | pass→pass | 4,254 | 4,665 | +10% | 1 | 1 | 0% | 828 | 1,474 | +78% | 0 | 0 | — |
case-03 | pass→pass | 3,788 | 4,793 | +27% | 1 | 1 | 0% | 759 | 1,412 | +86% | 0 | 0 | — |
case-04 | pass→pass | 4,723 | 4,268 | -10% | 1 | 1 | 0% | 876 | 1,358 | +55% | 0 | 0 | — |
case-05 | pass→pass | 3,737 | 5,836 | +56% | 1 | 1 | 0% | 658 | 1,500 | +128% | 0 | 0 | — |
case-06 | pass→pass | 3,951 | 4,994 | +26% | 1 | 1 | 0% | 769 | 1,417 | +84% | 0 | 0 | — |
case-07 | pass→pass | 4,557 | 6,956 | +53% | 1 | 1 | 0% | 746 | 1,736 | +133% | 0 | 0 | — |
case-08 | pass→pass | 9,245 | 12,822 | +39% | 1 | 1 | 0% | 1,710 | 2,859 | +67% | 0 | 0 | — |
case-09 | pass→pass | 11,633 | 15,798 | +36% | 1 | 1 | 0% | 2,237 | 3,496 | +56% | 0 | 0 | — |
case-10 | pass→pass | 4,502 | 6,642 | +48% | 1 | 1 | 0% | 958 | 1,727 | +80% | 0 | 0 | — |
case-11 | pass→pass | 11,693 | 11,491 | -2% | 1 | 1 | 0% | 2,228 | 2,594 | +16% | 0 | 0 | — |
case-12 | pass→pass | 15,713 | 15,065 | -4% | 1 | 1 | 0% | 2,588 | 3,151 | +22% | 0 | 0 | — |
case-13 | pass→pass | 16,827 | 10,577 | -37% | 1 | 1 | 0% | 1,967 | 2,144 | +9% | 0 | 0 | — |
case-14 | pass→pass | 15,928 | 20,927 | +31% | 1 | 1 | 0% | 2,325 | 2,539 | +9% | 0 | 0 | — |
case-15 | pass→pass | 10,634 | 10,510 | -1% | 1 | 1 | 0% | 986 | 1,635 | +66% | 0 | 0 | — |
case-16 | pass→pass | 21,283 | 13,267 | -38% | 1 | 1 | 0% | 1,660 | 3,168 | +91% | 0 | 0 | — |
case-17 | pass→pass | 16,385 | 16,330 | -0% | 1 | 1 | 0% | 1,883 | 2,632 | +40% | 0 | 0 | — |
case-18 | fail→pass | 19,593 | 11,301 | -42% | 1 | 1 | 0% | 1,881 | 2,375 | +26% | 0 | 0 | — |
case-19 | pass→pass | 1,654 | 2,111 | +28% | 1 | 1 | 0% | 283 | 993 | +251% | 0 | 0 | — |
case-20 | pass→pass | 15,519 | 11,072 | -29% | 1 | 1 | 0% | 3,072 | 2,517 | -18% | 0 | 0 | — |
case-21 | pass→pass | 15,191 | 14,901 | -2% | 1 | 1 | 0% | 2,888 | 3,486 | +21% | 0 | 0 | — |
case-22 | pass→pass | 4,123 | 6,002 | +46% | 1 | 1 | 0% | 875 | 1,713 | +96% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +5 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.