Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a home workout plan that fits your gear, time, and goal — a real weekly structure, not a random list of exercises. Use when asked to build a home workout, make me a workout plan, exercise routine at home, or how do I work out with no gym. Produces a weekly plan matched to your equipment and schedule, each session with warm-up, main work, sets/reps, and progression, swaps for missing gear, and a way to make it harder over time — with a plain 'this isn't medical advice, stop if it hurts' not
.claude/skills/mohitagw15856-home-workout-builder/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 35% | 0% |
A pile of random exercises isn't a program. This builds a structured weekly plan around what you actually have — bodyweight, a couple of dumbbells, a band — and where you want to go, with a way to progress so you keep getting results instead of plateauing or losing interest.
Ask for these if not provided:
Weekly structure: e.g. Mon full-body, Wed …].
Session A — focus]
Progression: week to week — add reps/sets/tempo/harder variation].
> Not medical advice. Use good form, stop if something hurts (not the same as muscle burn), and check with a professional for injuries or health conditions.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 34,521 | 32,111 | -7% | 1 | 1 | 0% | 4,325 | 5,436 | +26% | 0 | 0 | — |
case-02 | fail→pass | 40,697 | 27,557 | -32% | 1 | 1 | 0% | 6,526 | 4,690 | -28% | 0 | 0 | — |
case-03 | fail→pass | 30,627 | 25,505 | -17% | 1 | 1 | 0% | 4,496 | 4,409 | -2% | 0 | 0 | — |
case-04 | pass→pass | 19,792 | 21,124 | +7% | 1 | 1 | 0% | 2,177 | 3,171 | +46% | 0 | 0 | — |
case-05 | fail→fail | 28,027 | 32,843 | +17% | 1 | 1 | 0% | 4,420 | 6,168 | +40% | 0 | 0 | — |
case-06 | fail→fail | 38,687 | 24,391 | -37% | 1 | 1 | 0% | 5,684 | 4,256 | -25% | 0 | 0 | — |
case-07 | pass→pass | 14,127 | 10,429 | -26% | 1 | 1 | 0% | 1,602 | 1,681 | +5% | 0 | 0 | — |
case-08 | fail→fail | 18,195 | 13,298 | -27% | 1 | 1 | 0% | 2,663 | 3,439 | +29% | 0 | 0 | — |
case-09 | fail→pass | 26,229 | 18,298 | -30% | 1 | 1 | 0% | 3,605 | 3,897 | +8% | 0 | 0 | — |
case-10 | fail→pass | 22,970 | 17,949 | -22% | 1 | 1 | 0% | 3,095 | 4,190 | +35% | 0 | 0 | — |
case-11 | pass→pass | 15,383 | 17,364 | +13% | 1 | 1 | 0% | 2,871 | 4,204 | +46% | 0 | 0 | — |
case-12 | fail→pass | 21,767 | 24,566 | +13% | 1 | 1 | 0% | 3,865 | 4,473 | +16% | 0 | 0 | — |
case-13 | pass→pass | 22,624 | 21,263 | -6% | 1 | 1 | 0% | 3,313 | 3,888 | +17% | 0 | 0 | — |
case-14 | pass→pass | 20,116 | 22,527 | +12% | 1 | 1 | 0% | 3,761 | 3,997 | +6% | 0 | 0 | — |
case-15 | pass→pass | 19,978 | 21,914 | +10% | 1 | 1 | 0% | 2,778 | 4,065 | +46% | 0 | 0 | — |
case-16 | fail→pass | 22,845 | 18,319 | -20% | 1 | 1 | 0% | 3,047 | 3,995 | +31% | 0 | 0 | — |
case-17 | fail→pass | 22,407 | 20,933 | -7% | 1 | 1 | 0% | 3,277 | 4,276 | +30% | 0 | 0 | — |
case-18 | fail→pass | 28,121 | 24,430 | -13% | 1 | 1 | 0% | 4,924 | 5,452 | +11% | 0 | 0 | — |
case-19 | pass→pass | 23,935 | 22,142 | -7% | 1 | 1 | 0% | 3,442 | 4,341 | +26% | 0 | 0 | — |
case-20 | pass→pass | 23,266 | 17,306 | -26% | 1 | 1 | 0% | 3,484 | 3,920 | +13% | 0 | 0 | — |
case-21 | fail→pass | 21,005 | 22,100 | +5% | 1 | 1 | 0% | 2,990 | 4,092 | +37% | 0 | 0 | — |
case-22 | fail→pass | 20,828 | 24,670 | +18% | 1 | 1 | 0% | 3,256 | 3,856 | +18% | 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 +50 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.