Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Check whether an online store or seller is legit before you pay — and pay in a way you can get your money back if it isn't. Use when asked is this website legit, is this online store a scam, should I buy from this site, or how to shop safely online. Produces a trust assessment from the store's signals (too-good pricing, contact/policy gaps, domain and review red flags), safe-payment guidance that preserves buyer protection, what to check before checkout, and what to do if you've already paid a s
.claude/skills/mohitagw15856-safe-online-shopping/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-05 | ✓→✗ | ▼ Worse | 129% | 0% |
| case-01 | ✓→✓ | = Same ✓ | -7% | 0% |
| case-02 | ✓→✓ | = Same ✓ | -5% | 0% |
Fake shops are everywhere: a slick site, an unbeatable price, and no way to get your money back once you've paid by bank transfer. This checks the store's trust signals before you buy, steers you to a payment method with buyer protection, and — if you've already paid a site that now looks wrong — switches to getting your money back.
Ask for these if not provided:
Trust signals
Before checkout: verify contact/policies · check domain/reviews · confirm secure payment]. Pay with: credit card / protected processor] — avoid transfer/gift card/crypto]. Verdict: proceed / careful / avoid] — because x].
Already paid and it's a scam? card chargeback/dispute · gather evidence · report the site].
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 20,906 | 15,152 | -28% | 1 | 1 | 0% | 2,856 | 2,643 | -7% | 0 | 0 | — |
case-02 | pass→pass | 26,188 | 15,175 | -42% | 1 | 1 | 0% | 3,009 | 2,865 | -5% | 0 | 0 | — |
case-03 | pass→pass | 21,554 | 16,538 | -23% | 1 | 1 | 0% | 2,648 | 3,133 | +18% | 0 | 0 | — |
case-04 | pass→pass | 21,014 | 10,998 | -48% | 1 | 1 | 0% | 2,251 | 2,720 | +21% | 0 | 0 | — |
case-05 | pass→fail | 6,257 | 10,123 | +62% | 1 | 1 | 0% | 1,106 | 2,536 | +129% | 0 | 0 | — |
case-06 | fail→pass | 16,059 | 10,434 | -35% | 1 | 1 | 0% | 1,655 | 2,353 | +42% | 0 | 0 | — |
case-07 | pass→pass | 15,517 | 13,382 | -14% | 1 | 1 | 0% | 1,667 | 2,400 | +44% | 0 | 0 | — |
case-08 | pass→pass | 20,247 | 10,860 | -46% | 1 | 1 | 0% | 2,006 | 2,699 | +35% | 0 | 0 | — |
case-09 | pass→pass | 15,002 | 10,692 | -29% | 1 | 1 | 0% | 1,616 | 2,018 | +25% | 0 | 0 | — |
case-10 | pass→pass | 14,434 | 13,170 | -9% | 1 | 1 | 0% | 2,131 | 2,284 | +7% | 0 | 0 | — |
case-11 | pass→pass | 17,883 | 13,103 | -27% | 1 | 1 | 0% | 1,785 | 2,336 | +31% | 0 | 0 | — |
case-12 | pass→pass | 19,588 | 13,198 | -33% | 1 | 1 | 0% | 1,950 | 2,359 | +21% | 0 | 0 | — |
case-13 | pass→pass | 14,945 | 9,031 | -40% | 1 | 1 | 0% | 1,633 | 2,419 | +48% | 0 | 0 | — |
case-14 | pass→pass | 15,815 | 13,588 | -14% | 1 | 1 | 0% | 1,743 | 2,323 | +33% | 0 | 0 | — |
case-15 | pass→pass | 17,433 | 8,795 | -50% | 1 | 1 | 0% | 1,923 | 2,350 | +22% | 0 | 0 | — |
case-16 | fail→pass | 22,283 | 10,419 | -53% | 1 | 1 | 0% | 2,229 | 2,710 | +22% | 0 | 0 | — |
case-17 | pass→pass | 18,181 | 15,991 | -12% | 1 | 1 | 0% | 2,102 | 2,679 | +27% | 0 | 0 | — |
case-18 | pass→pass | 17,946 | 16,836 | -6% | 1 | 1 | 0% | 2,107 | 2,887 | +37% | 0 | 0 | — |
case-19 | pass→pass | 15,932 | 13,854 | -13% | 1 | 1 | 0% | 1,738 | 2,390 | +38% | 0 | 0 | — |
case-20 | pass→pass | 14,199 | 9,710 | -32% | 1 | 1 | 0% | 1,408 | 2,206 | +57% | 0 | 0 | — |
case-21 | pass→pass | 11,595 | 13,692 | +18% | 1 | 1 | 0% | 1,476 | 2,314 | +57% | 0 | 0 | — |
case-22 | pass→pass | 12,866 | 14,254 | +11% | 1 | 1 | 0% | 1,188 | 2,185 | +84% | 0 | 0 | — |
case-23 | pass→pass | 14,561 | 12,372 | -15% | 1 | 1 | 0% | 1,489 | 2,156 | +45% | 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.