Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a realistic plan to fix bad sleep — a wind-down routine, a consistent schedule, and the daytime and environment fixes that actually move the needle. Use when asked to fix my sleep, I can't sleep, help me sleep better, or build a bedtime routine. Produces a read on the likely disruptors, a wind-down sequence, schedule and light/caffeine timing, environment tweaks, and a 'what to do when you can't fall asleep' rule — flagging that persistent insomnia or symptoms like snoring/apnea warrant a
.claude/skills/mohitagw15856-sleep-reset-plan/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-07 | ✓→✗ | ▼ Worse | 32% | 0% |
| case-06 | ✓→✓ | = Same ✓ | 67% | 0% |
| case-14 | ✓→✓ | = Same ✓ | 39% | 0% |
Good sleep is mostly built during the day and the hour before bed — not willed at 2am. This looks at what's likely wrecking your sleep (screens, caffeine timing, an erratic schedule, a too-warm room), and builds a realistic wind-down and daytime routine, plus a rule for the nights you're staring at the ceiling — while being clear that ongoing insomnia or breathing issues need a doctor.
Ask for these if not provided:
Likely disruptors: top 2–3 from habits].
Wind-down (last 30–60 min): step → step → step]. Schedule & timing: wake consistent time] · morning light · caffeine cutoff time] · alcohol/naps rule]. Environment: dark · cool · quiet · phone out of reach. Can't sleep after ~20 min? get up, dim, calm activity, return when sleepy].
> See a doctor if insomnia persists for weeks, or if there's loud snoring/gasping or heavy daytime sleepiness (possible sleep apnea).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 18,022 | 22,863 | +27% | 1 | 1 | 0% | 1,972 | 3,300 | +67% | 0 | 0 | — |
case-14 | pass→pass | 20,683 | 19,411 | -6% | 1 | 1 | 0% | 2,167 | 3,006 | +39% | 0 | 0 | — |
case-05 | pass→pass | 15,816 | 12,697 | -20% | 1 | 1 | 0% | 1,713 | 3,064 | +79% | 0 | 0 | — |
case-01 | fail→pass | 26,542 | 16,559 | -38% | 1 | 1 | 0% | 3,709 | 3,101 | -16% | 0 | 0 | — |
case-02 | fail→fail | 26,578 | 11,938 | -55% | 1 | 1 | 0% | 3,878 | 3,065 | -21% | 0 | 0 | — |
case-03 | fail→pass | 26,767 | 17,353 | -35% | 1 | 1 | 0% | 3,359 | 3,590 | +7% | 0 | 0 | — |
case-04 | pass→pass | 16,720 | 20,556 | +23% | 1 | 1 | 0% | 1,798 | 3,005 | +67% | 0 | 0 | — |
case-07 | pass→fail | 21,106 | 18,205 | -14% | 1 | 1 | 0% | 2,349 | 3,103 | +32% | 0 | 0 | — |
case-08 | pass→pass | 17,570 | 17,975 | +2% | 1 | 1 | 0% | 2,063 | 3,209 | +56% | 0 | 0 | — |
case-09 | pass→pass | 15,509 | 16,940 | +9% | 1 | 1 | 0% | 1,784 | 2,723 | +53% | 0 | 0 | — |
case-10 | pass→pass | 21,838 | 11,525 | -47% | 1 | 1 | 0% | 2,275 | 2,892 | +27% | 0 | 0 | — |
case-11 | pass→pass | 18,928 | 18,569 | -2% | 1 | 1 | 0% | 1,993 | 2,827 | +42% | 0 | 0 | — |
case-12 | pass→pass | 108,000 | 17,976 | -83% | 1 | 1 | 0% | 2,041 | 3,056 | +50% | 0 | 0 | — |
case-13 | pass→pass | 11,187 | 21,228 | +90% | 1 | 1 | 0% | 1,556 | 3,223 | +107% | 0 | 0 | — |
case-15 | pass→pass | 15,820 | 19,569 | +24% | 1 | 1 | 0% | 2,347 | 2,981 | +27% | 0 | 0 | — |
case-16 | pass→pass | 16,572 | 17,717 | +7% | 1 | 1 | 0% | 1,781 | 2,576 | +45% | 0 | 0 | — |
case-17 | pass→pass | 18,429 | 15,993 | -13% | 1 | 1 | 0% | 1,681 | 2,347 | +40% | 0 | 0 | — |
case-18 | pass→pass | 15,675 | 19,486 | +24% | 1 | 1 | 0% | 2,425 | 2,932 | +21% | 0 | 0 | — |
case-19 | pass→pass | 21,311 | 18,359 | -14% | 1 | 1 | 0% | 3,072 | 3,068 | -0% | 0 | 0 | — |
case-20 | pass→pass | 18,659 | 21,882 | +17% | 1 | 1 | 0% | 2,158 | 3,136 | +45% | 0 | 0 | — |
case-21 | pass→pass | 24,382 | 23,749 | -3% | 1 | 1 | 0% | 2,410 | 3,506 | +45% | 0 | 0 | — |
case-22 | pass→pass | 21,321 | 19,948 | -6% | 1 | 1 | 0% | 2,148 | 3,206 | +49% | 0 | 0 | — |
case-23 | pass→pass | 21,995 | 16,943 | -23% | 1 | 1 | 0% | 2,175 | 2,703 | +24% | 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. 23 cases were attempted. The headline lift of +4 percentage points is the difference between those two pass rates over the 23 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.