Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Decide what to cook tonight from what you already have and how much time/energy you've got — no shopping trip, no recipe rabbit hole. Use when asked what should I make for dinner, what can I cook with what's in my fridge, quick dinner ideas, or I don't know what to eat. Produces 3 doable options ranked by effort with a quick method for each, honest substitutions, and a 'need one thing' flag if a near-miss is worth a corner-shop run — respecting diets and dislikes.
.claude/skills/mohitagw15856-whats-for-dinner/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-11 | ✓→✗ | ▼ Worse | 21% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -18% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 20% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 20% | 0% |
The nightly stall isn't a lack of recipes — it's decision fatigue plus a fridge of odd ingredients. This looks at what you actually have, how much effort you have in you tonight, and any diet lines, then gives three real options ranked by effort with just enough method to cook — no 900-word life story, no assuming you'll pop out for five ingredients.
Ask for these if not provided:
🥇 Dish] — ~mins], effort] Uses: … · Method: 1) … 2) … 3) … · Sub: if missing X, use Y]
🥈 Dish] — ~mins] …
🥉 Dish] — ~mins] …
Need one thing? Only if a standout is one cheap item away] — otherwise skip.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 21,597 | 14,709 | -32% | 1 | 1 | 0% | 3,065 | 2,517 | -18% | 0 | 0 | — |
case-02 | pass→pass | 21,820 | 20,311 | -7% | 1 | 1 | 0% | 2,862 | 3,439 | +20% | 0 | 0 | — |
case-03 | fail→pass | 30,276 | 21,552 | -29% | 1 | 1 | 0% | 4,363 | 3,565 | -18% | 0 | 0 | — |
case-04 | pass→pass | 15,297 | 19,328 | +26% | 1 | 1 | 0% | 2,674 | 3,202 | +20% | 0 | 0 | — |
case-05 | pass→pass | 11,872 | 19,672 | +66% | 1 | 1 | 0% | 1,819 | 3,277 | +80% | 0 | 0 | — |
case-06 | pass→pass | 17,886 | 15,883 | -11% | 1 | 1 | 0% | 2,167 | 2,647 | +22% | 0 | 0 | — |
case-07 | pass→pass | 14,446 | 18,723 | +30% | 1 | 1 | 0% | 2,377 | 2,965 | +25% | 0 | 0 | — |
case-08 | pass→pass | 20,974 | 13,054 | -38% | 1 | 1 | 0% | 2,674 | 3,079 | +15% | 0 | 0 | — |
case-09 | pass→pass | 17,412 | 12,623 | -28% | 1 | 1 | 0% | 2,172 | 2,928 | +35% | 0 | 0 | — |
case-10 | pass→pass | 13,940 | 18,082 | +30% | 1 | 1 | 0% | 2,346 | 2,970 | +27% | 0 | 0 | — |
case-11 | pass→fail | 19,462 | 13,086 | -33% | 1 | 1 | 0% | 2,665 | 3,215 | +21% | 0 | 0 | — |
case-12 | pass→pass | 18,041 | 18,455 | +2% | 1 | 1 | 0% | 2,212 | 3,139 | +42% | 0 | 0 | — |
case-13 | pass→pass | 19,574 | 16,557 | -15% | 1 | 1 | 0% | 2,403 | 2,692 | +12% | 0 | 0 | — |
case-14 | pass→pass | 15,591 | 18,605 | +19% | 1 | 1 | 0% | 2,673 | 3,107 | +16% | 0 | 0 | — |
case-15 | pass→pass | 19,343 | 15,200 | -21% | 1 | 1 | 0% | 2,333 | 2,596 | +11% | 0 | 0 | — |
case-16 | pass→pass | 14,360 | 11,577 | -19% | 1 | 1 | 0% | 2,430 | 2,831 | +17% | 0 | 0 | — |
case-17 | pass→pass | 15,619 | 14,090 | -10% | 1 | 1 | 0% | 2,102 | 3,043 | +45% | 0 | 0 | — |
case-18 | pass→pass | 11,399 | 19,191 | +68% | 1 | 1 | 0% | 1,880 | 3,232 | +72% | 0 | 0 | — |
case-19 | pass→pass | 23,150 | 14,880 | -36% | 1 | 1 | 0% | 3,338 | 3,448 | +3% | 0 | 0 | — |
case-20 | pass→pass | 27,466 | 24,072 | -12% | 1 | 1 | 0% | 3,687 | 4,880 | +32% | 0 | 0 | — |
case-21 | pass→pass | 35,199 | 17,674 | -50% | 1 | 1 | 0% | 4,283 | 3,959 | -8% | 0 | 0 | — |
case-22 | pass→pass | 16,660 | 13,917 | -16% | 1 | 1 | 0% | 3,114 | 2,907 | -7% | 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 0 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.