Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Plan a genuine break from caregiving — arrange the coverage, hand off the essentials, and actually rest — because respite is what lets you keep going. Use when asked I need a break from caregiving, how do I arrange respite care, I can't leave them alone, or help me get time off from caring. Produces the coverage options for your situation (family, paid respite, day programs, short-stay), a handoff pack so whoever covers has what they need, how to overcome the barriers (guilt, trust, cost, logist
.claude/skills/mohitagw15856-respite-care-plan/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 147% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 79% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 112% | 0% |
"I can't leave them" is how caregivers burn out — but respite (planned breaks) is exactly what makes long-term caregiving sustainable. The barriers are real: guilt, not trusting anyone else, cost, and the logistics of handing off. This plans a genuine break: who can cover, the handoff so they're set up to succeed, how to get past the barriers, and how to actually rest instead of worrying the whole time — so the break happens and repeats. Not medical advice.
Ask for these if not provided:
Coverage options: family/friends · paid in-home respite · day program · short-stay] — with pros/cons for you. Handoff pack (so you can step away): routine · medications · contacts · warning signs · preferences. Bust the barriers: guilt → necessary, not selfish] · trust → prepared stand-in] · cost → funding to research] · logistics → start small]. Actually rest: how to genuinely recover + quiet the worry — a check-in, trusting the pack]. Make it repeatable: a regular rhythm of breaks, not a one-off].
> Logistics support — not medical advice. For medically complex care, arrange professional respite services.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 30,396 | 27,847 | -8% | 1 | 1 | 0% | 3,500 | 4,239 | +21% | 0 | 0 | — |
case-02 | pass→pass | 29,785 | 22,737 | -24% | 1 | 1 | 0% | 3,921 | 3,924 | +0% | 0 | 0 | — |
case-03 | fail→pass | 33,859 | 25,001 | -26% | 1 | 1 | 0% | 4,719 | 4,425 | -6% | 0 | 0 | — |
case-04 | pass→pass | 13,181 | 19,266 | +46% | 1 | 1 | 0% | 1,472 | 3,741 | +154% | 0 | 0 | — |
case-05 | pass→pass | 18,545 | 23,609 | +27% | 1 | 1 | 0% | 2,278 | 4,041 | +77% | 0 | 0 | — |
case-06 | pass→pass | 22,624 | 19,707 | -13% | 1 | 1 | 0% | 2,284 | 3,459 | +51% | 0 | 0 | — |
case-07 | fail→pass | 16,610 | 23,647 | +42% | 1 | 1 | 0% | 1,634 | 4,033 | +147% | 0 | 0 | — |
case-08 | pass→pass | 14,428 | 13,588 | -6% | 1 | 1 | 0% | 2,272 | 3,229 | +42% | 0 | 0 | — |
case-09 | pass→pass | 20,545 | 25,352 | +23% | 1 | 1 | 0% | 2,389 | 3,843 | +61% | 0 | 0 | — |
case-10 | pass→pass | 19,689 | 21,700 | +10% | 1 | 1 | 0% | 2,188 | 3,886 | +78% | 0 | 0 | — |
case-11 | fail→pass | 17,151 | 14,334 | -16% | 1 | 1 | 0% | 1,893 | 3,389 | +79% | 0 | 0 | — |
case-12 | pass→pass | 22,345 | 17,139 | -23% | 1 | 1 | 0% | 2,367 | 3,727 | +57% | 0 | 0 | — |
case-13 | pass→fail | 18,443 | 21,001 | +14% | 1 | 1 | 0% | 2,814 | 3,584 | +27% | 0 | 0 | — |
case-14 | pass→pass | 21,048 | 19,071 | -9% | 1 | 1 | 0% | 2,327 | 3,397 | +46% | 0 | 0 | — |
case-15 | pass→pass | 18,637 | 21,208 | +14% | 1 | 1 | 0% | 2,109 | 3,633 | +72% | 0 | 0 | — |
case-16 | fail→pass | 17,243 | 22,102 | +28% | 1 | 1 | 0% | 1,712 | 3,633 | +112% | 0 | 0 | — |
case-17 | pass→pass | 22,231 | 17,580 | -21% | 1 | 1 | 0% | 2,279 | 3,913 | +72% | 0 | 0 | — |
case-18 | pass→pass | 17,168 | 23,000 | +34% | 1 | 1 | 0% | 2,544 | 3,811 | +50% | 0 | 0 | — |
case-19 | pass→pass | 19,894 | 20,583 | +3% | 1 | 1 | 0% | 2,184 | 3,629 | +66% | 0 | 0 | — |
case-20 | fail→pass | 15,648 | 24,649 | +58% | 1 | 1 | 0% | 2,363 | 4,084 | +73% | 0 | 0 | — |
case-21 | fail→pass | 18,051 | 23,993 | +33% | 1 | 1 | 0% | 2,267 | 3,999 | +76% | 0 | 0 | — |
case-22 | fail→pass | 10,744 | 21,405 | +99% | 1 | 1 | 0% | 1,689 | 3,874 | +129% | 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 +32 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.