Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when reviewing image assets, markup, and CDN or build transforms related to Use WebP format with fallbacks. Check encoded size, rendered size, loading strategy, and above-the-fold impact together.
.claude/skills/thedaviddias-webp-format/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 4 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 25% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -39% | 0% |
WebP is smaller than JPEG and PNG at similar quality—using it reduces page weight by 25-35% for images, improving load times on all devices.
Check if images are served in modern formats like WebP with fallbacks for older browsers.
Convert images to WebP format and implement proper fallback strategies.
Explain how WebP provides 25-35% better compression than JPEG and PNG.
Review image assets, markup, and delivery configuration related to Use WebP format with fallbacks. Flag exact files or components where format choice, sizing, or loading behavior violates the rule, and describe how to confirm the fix in DevTools.
For full implementation details, code examples, and framework-specific guidance, see references/rule.md.
Rule page: https://frontendchecklist.io/en/rules/images/webp-format
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 12,960 | 9,494 | -27% | 1 | 1 | 0% | 2,502 | 2,143 | -14% | 0 | 0 | — |
case-02 | fail→fail | 16,627 | 13,195 | -21% | 1 | 1 | 0% | 2,725 | 2,424 | -11% | 0 | 0 | — |
case-03 | fail→pass | 7,982 | 4,285 | -46% | 1 | 1 | 0% | 1,448 | 1,018 | -30% | 0 | 0 | — |
case-04 | pass→pass | 4,151 | 4,005 | -4% | 1 | 1 | 0% | 754 | 944 | +25% | 0 | 0 | — |
case-05 | pass→pass | 13,281 | 7,084 | -47% | 1 | 1 | 0% | 2,465 | 1,506 | -39% | 0 | 0 | — |
case-06 | pass→pass | 7,268 | 6,166 | -15% | 1 | 1 | 0% | 1,354 | 1,312 | -3% | 0 | 0 | — |
case-07 | pass→pass | 10,331 | 6,889 | -33% | 1 | 1 | 0% | 1,759 | 1,469 | -16% | 0 | 0 | — |
case-08 | pass→pass | 10,153 | 5,808 | -43% | 1 | 1 | 0% | 1,825 | 1,276 | -30% | 0 | 0 | — |
case-09 | pass→pass | 10,649 | 3,529 | -67% | 1 | 1 | 0% | 1,755 | 810 | -54% | 0 | 0 | — |
case-10 | pass→pass | 5,607 | 3,969 | -29% | 1 | 1 | 0% | 1,070 | 951 | -11% | 0 | 0 | — |
case-11 | pass→pass | 14,420 | 13,090 | -9% | 1 | 1 | 0% | 2,854 | 2,657 | -7% | 0 | 0 | — |
case-12 | fail→pass | 8,149 | 5,124 | -37% | 1 | 1 | 0% | 1,649 | 1,173 | -29% | 0 | 0 | — |
case-13 | pass→pass | 11,706 | 6,954 | -41% | 1 | 1 | 0% | 1,882 | 1,446 | -23% | 0 | 0 | — |
case-14 | pass→pass | 13,626 | 8,714 | -36% | 1 | 1 | 0% | 2,247 | 1,817 | -19% | 0 | 0 | — |
case-15 | pass→pass | 11,671 | 6,538 | -44% | 1 | 1 | 0% | 1,916 | 1,398 | -27% | 0 | 0 | — |
case-16 | pass→pass | 13,626 | 12,056 | -12% | 1 | 1 | 0% | 2,697 | 2,569 | -5% | 0 | 0 | — |
case-17 | pass→pass | 10,418 | 43,390 | +316% | 1 | 1 | 0% | 1,964 | 1,376 | -30% | 0 | 0 | — |
case-18 | pass→pass | 6,308 | 6,084 | -4% | 1 | 1 | 0% | 1,134 | 1,334 | +18% | 0 | 0 | — |
case-19 | pass→pass | 9,679 | 7,724 | -20% | 1 | 1 | 0% | 1,781 | 1,734 | -3% | 0 | 0 | — |
case-20 | pass→pass | 15,984 | 11,678 | -27% | 1 | 1 | 0% | 2,740 | 2,276 | -17% | 0 | 0 | — |
case-21 | pass→pass | 11,371 | 6,931 | -39% | 1 | 1 | 0% | 2,106 | 1,534 | -27% | 0 | 0 | — |
case-22 | pass→pass | 13,860 | 8,306 | -40% | 1 | 1 | 0% | 2,506 | 1,923 | -23% | 0 | 0 | — |
case-23 | pass→pass | 9,037 | 4,883 | -46% | 1 | 1 | 0% | 1,617 | 1,182 | -27% | 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.