Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Find out whether something you own — a car, appliance, car seat, food item, or gadget — is under a safety recall, and what to do about it. Use when asked to check for a recall, is my [product] recalled, I heard about a recall on, or how do I find out if my car/appliance is affected. Produces a structured way to check by make/model/batch against the official sources, how to read whether your specific unit is affected, the free remedy you're owed, urgency triage for safety risks, and how to regist
.claude/skills/mohitagw15856-product-recall-check/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 69% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 77% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 68% | 0% |
| case-01 | ✓→✓ | = Same ✓ | 21% | 0% |
| case-02 | ✓→✓ | = Same ✓ | 16% | 0% |
Recalls are the rare case where a company is legally required to fix or replace your thing for free — but only if you know it's recalled and act. Notices get missed: you bought secondhand, moved house, or never registered. This gives you a reliable way to check the right official sources by the details that matter (make, model, serial/VIN, batch, date), read whether your specific unit is in scope, and claim the free remedy — with honest urgency for anything that's a genuine safety risk.
Ask for these if not provided:
Always confirm against the current official database — recalls are added and updated constantly, and this can't stand in for the live source.
Where to check: the right official register for this product type] — search by VIN / serial / model / batch].
Is yours affected? Match your identifier] against the recall's stated range: how to read it]. Recalls often cover only specific runs — confirm the exact numbers.
If it IS recalled
Set up alerts: register the product with the maker + subscribe to official recall notifications for product type].
> Confirm everything against the current official recall database — this is a guide to checking, not a substitute for the live source.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 26,357 | 31,862 | +21% | 1 | 1 | 0% | 2,803 | 3,388 | +21% | 0 | 0 | — |
case-02 | pass→pass | 27,065 | 22,987 | -15% | 1 | 1 | 0% | 2,967 | 3,448 | +16% | 0 | 0 | — |
case-03 | pass→pass | 17,493 | 14,783 | -15% | 1 | 1 | 0% | 2,037 | 3,206 | +57% | 0 | 0 | — |
case-04 | pass→pass | 12,010 | 20,096 | +67% | 1 | 1 | 0% | 2,030 | 3,131 | +54% | 0 | 0 | — |
case-05 | pass→pass | 13,718 | 15,371 | +12% | 1 | 1 | 0% | 1,366 | 2,786 | +104% | 0 | 0 | — |
case-06 | pass→pass | 16,451 | 21,444 | +30% | 1 | 1 | 0% | 1,534 | 3,303 | +115% | 0 | 0 | — |
case-07 | pass→pass | 15,507 | 21,183 | +37% | 1 | 1 | 0% | 1,541 | 3,163 | +105% | 0 | 0 | — |
case-08 | fail→pass | 17,079 | 17,670 | +3% | 1 | 1 | 0% | 1,788 | 3,018 | +69% | 0 | 0 | — |
case-09 | pass→pass | 14,916 | 15,868 | +6% | 1 | 1 | 0% | 1,526 | 2,753 | +80% | 0 | 0 | — |
case-10 | pass→pass | 17,742 | 19,279 | +9% | 1 | 1 | 0% | 1,631 | 3,053 | +87% | 0 | 0 | — |
case-11 | pass→pass | 16,252 | 13,132 | -19% | 1 | 1 | 0% | 1,675 | 2,902 | +73% | 0 | 0 | — |
case-12 | fail→pass | 16,936 | 16,151 | -5% | 1 | 1 | 0% | 1,690 | 2,995 | +77% | 0 | 0 | — |
case-13 | pass→pass | 17,136 | 28,317 | +65% | 1 | 1 | 0% | 1,740 | 3,170 | +82% | 0 | 0 | — |
case-14 | pass→pass | 11,081 | 20,250 | +83% | 1 | 1 | 0% | 1,689 | 3,109 | +84% | 0 | 0 | — |
case-15 | pass→pass | 16,345 | 17,726 | +8% | 1 | 1 | 0% | 1,742 | 3,020 | +73% | 0 | 0 | — |
case-16 | fail→pass | 15,360 | 18,055 | +18% | 1 | 1 | 0% | 1,925 | 3,231 | +68% | 0 | 0 | — |
case-17 | pass→pass | 11,298 | 11,054 | -2% | 1 | 1 | 0% | 1,629 | 2,766 | +70% | 0 | 0 | — |
case-18 | pass→pass | 11,758 | 17,804 | +51% | 1 | 1 | 0% | 1,823 | 3,465 | +90% | 0 | 0 | — |
case-19 | pass→pass | 23,268 | 15,294 | -34% | 1 | 1 | 0% | 1,784 | 2,986 | +67% | 0 | 0 | — |
case-20 | pass→pass | 15,328 | 22,437 | +46% | 1 | 1 | 0% | 2,307 | 3,381 | +47% | 0 | 0 | — |
case-21 | pass→pass | 15,247 | 15,324 | +1% | 1 | 1 | 0% | 2,164 | 3,722 | +72% | 0 | 0 | — |
case-22 | pass→pass | 12,579 | 20,581 | +64% | 1 | 1 | 0% | 1,805 | 3,987 | +121% | 0 | 0 | — |
case-23 | pass→pass | 10,642 | 26,026 | +145% | 1 | 1 | 0% | 1,474 | 2,791 | +89% | 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 +13 percentage points is the difference between those two pass rates over the 23 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.