Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Do the right things fast when your phone is lost or stolen — lock it, protect your accounts and money, and decide on wipe vs. locate — in the correct order. Use when asked my phone was stolen, I lost my phone what do I do, someone took my phone, or secure my lost phone. Produces an ordered action checklist (locate/lock, protect SIM and banking, change key passwords, wipe decision), the accounts to prioritize because the phone unlocks them, reporting steps, and prevention setup for next time.
.claude/skills/mohitagw15856-secure-a-lost-phone/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 46% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 38% | 0% |
A lost phone isn't just a lost device — it's a key to your email, banking, 2FA, and payments. The first 30 minutes matter. This gives the correct order: locate and lock it, stop the SIM and card access, change the passwords that matter, and decide when to remotely wipe — so a bad day doesn't become account takeover and fraud.
Ask for these if not provided:
Do now (in order)
Set up for next time: passcode + biometrics · find-my on · encrypted backups · 2FA that isn't only this phone.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 31,885 | 21,286 | -33% | 1 | 1 | 0% | 3,521 | 3,190 | -9% | 0 | 0 | — |
case-02 | fail→fail | 28,279 | 18,261 | -35% | 1 | 1 | 0% | 3,269 | 2,941 | -10% | 0 | 0 | — |
case-03 | fail→fail | 23,956 | 23,972 | +0% | 1 | 1 | 0% | 2,952 | 3,565 | +21% | 0 | 0 | — |
case-04 | pass→pass | 19,251 | 19,679 | +2% | 1 | 1 | 0% | 2,044 | 2,820 | +38% | 0 | 0 | — |
case-05 | pass→pass | 20,535 | 16,822 | -18% | 1 | 1 | 0% | 2,106 | 2,567 | +22% | 0 | 0 | — |
case-06 | fail→pass | 17,193 | 18,361 | +7% | 1 | 1 | 0% | 1,902 | 2,758 | +45% | 0 | 0 | — |
case-07 | pass→pass | 17,942 | 15,715 | -12% | 1 | 1 | 0% | 2,002 | 2,649 | +32% | 0 | 0 | — |
case-08 | pass→pass | 12,022 | 9,335 | -22% | 1 | 1 | 0% | 1,834 | 2,439 | +33% | 0 | 0 | — |
case-09 | pass→pass | 15,888 | 20,192 | +27% | 1 | 1 | 0% | 2,474 | 3,008 | +22% | 0 | 0 | — |
case-10 | pass→pass | 19,809 | 18,254 | -8% | 1 | 1 | 0% | 2,027 | 3,040 | +50% | 0 | 0 | — |
case-11 | pass→pass | 15,226 | 10,368 | -32% | 1 | 1 | 0% | 1,601 | 2,335 | +46% | 0 | 0 | — |
case-12 | fail→fail | 15,726 | 13,282 | -16% | 1 | 1 | 0% | 2,235 | 2,528 | +13% | 0 | 0 | — |
case-13 | fail→pass | 18,555 | 17,688 | -5% | 1 | 1 | 0% | 1,881 | 2,747 | +46% | 0 | 0 | — |
case-14 | fail→pass | 16,587 | 8,571 | -48% | 1 | 1 | 0% | 1,438 | 1,471 | +2% | 0 | 0 | — |
case-15 | pass→pass | 16,889 | 18,505 | +10% | 1 | 1 | 0% | 1,881 | 3,056 | +62% | 0 | 0 | — |
case-16 | fail→pass | 18,999 | 17,663 | -7% | 1 | 1 | 0% | 2,128 | 2,938 | +38% | 0 | 0 | — |
case-17 | pass→pass | 20,198 | 22,913 | +13% | 1 | 1 | 0% | 2,396 | 3,602 | +50% | 0 | 0 | — |
case-18 | pass→pass | 12,180 | 18,921 | +55% | 1 | 1 | 0% | 1,850 | 2,819 | +52% | 0 | 0 | — |
case-19 | pass→pass | 18,827 | 14,404 | -23% | 1 | 1 | 0% | 1,882 | 2,344 | +25% | 0 | 0 | — |
case-20 | pass→pass | 20,478 | 15,896 | -22% | 1 | 1 | 0% | 2,061 | 2,954 | +43% | 0 | 0 | — |
case-21 | fail→fail | 15,749 | 11,634 | -26% | 1 | 1 | 0% | 1,805 | 2,660 | +47% | 0 | 0 | — |
case-22 | pass→pass | 19,153 | 21,597 | +13% | 1 | 1 | 0% | 2,192 | 3,278 | +50% | 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 +18 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.