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 Serve images from a CDN. Check encoded size, rendered size, loading strategy, and above-the-fold impact together.
.claude/skills/thedaviddias-image-cdn/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-18 | ✓→✓ | ＝ Same ✓ | -18% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -8% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -18% | 0% |
Image CDNs automatically optimize, resize, and convert images on-the-fly—delivering the smallest file in the best format without manual processing.
Verify if images are served from a CDN or image optimization service.
Configure an image CDN like Cloudinary, Imgix, or Cloudflare Images for automatic optimization.
Explain how image CDNs provide on-the-fly optimization, resizing, and format conversion.
Review image assets, markup, and delivery configuration related to Serve images from a CDN. 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/image-cdn
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-18 | pass→pass | 9,668 | 7,451 | -23% | 1 | 1 | 0% | 2,111 | 1,734 | -18% | 0 | 0 | — |
case-04 | pass→pass | 11,533 | 10,177 | -12% | 1 | 1 | 0% | 2,422 | 2,218 | -8% | 0 | 0 | — |
case-01 | fail→pass | 12,448 | 12,504 | +0% | 1 | 1 | 0% | 2,887 | 2,782 | -4% | 0 | 0 | — |
case-02 | pass→pass | 12,558 | 8,993 | -28% | 1 | 1 | 0% | 2,651 | 2,169 | -18% | 0 | 0 | — |
case-03 | pass→pass | 14,151 | 11,321 | -20% | 1 | 1 | 0% | 2,766 | 2,587 | -6% | 0 | 0 | — |
case-05 | pass→pass | 12,558 | 9,013 | -28% | 1 | 1 | 0% | 2,426 | 1,997 | -18% | 0 | 0 | — |
case-06 | pass→pass | 11,452 | 8,215 | -28% | 1 | 1 | 0% | 2,327 | 1,688 | -27% | 0 | 0 | — |
case-07 | pass→pass | 11,751 | 6,888 | -41% | 1 | 1 | 0% | 2,207 | 1,881 | -15% | 0 | 0 | — |
case-08 | pass→pass | 12,781 | 8,822 | -31% | 1 | 1 | 0% | 2,220 | 1,816 | -18% | 0 | 0 | — |
case-09 | pass→pass | 9,738 | 5,139 | -47% | 1 | 1 | 0% | 1,615 | 1,245 | -23% | 0 | 0 | — |
case-10 | pass→pass | 11,926 | 7,887 | -34% | 1 | 1 | 0% | 2,002 | 1,650 | -18% | 0 | 0 | — |
case-11 | pass→pass | 3,005 | 2,592 | -14% | 1 | 1 | 0% | 537 | 789 | +47% | 0 | 0 | — |
case-12 | pass→pass | 13,625 | 9,651 | -29% | 1 | 1 | 0% | 2,418 | 2,338 | -3% | 0 | 0 | — |
case-13 | pass→pass | 14,757 | 8,565 | -42% | 1 | 1 | 0% | 2,871 | 2,274 | -21% | 0 | 0 | — |
case-14 | pass→pass | 18,790 | 7,492 | -60% | 1 | 1 | 0% | 2,467 | 1,769 | -28% | 0 | 0 | — |
case-15 | pass→pass | 11,790 | 7,024 | -40% | 1 | 1 | 0% | 2,445 | 1,664 | -32% | 0 | 0 | — |
case-16 | pass→pass | 10,011 | 5,885 | -41% | 1 | 1 | 0% | 2,009 | 1,336 | -33% | 0 | 0 | — |
case-17 | fail→pass | 7,901 | 7,337 | -7% | 1 | 1 | 0% | 1,802 | 1,912 | +6% | 0 | 0 | — |
case-19 | pass→pass | 12,060 | 6,564 | -46% | 1 | 1 | 0% | 2,118 | 1,744 | -18% | 0 | 0 | — |
case-20 | pass→pass | 15,450 | 8,865 | -43% | 1 | 1 | 0% | 2,600 | 2,058 | -21% | 0 | 0 | — |
case-21 | pass→pass | 18,058 | 7,361 | -59% | 1 | 1 | 0% | 2,343 | 1,521 | -35% | 0 | 0 | — |
case-22 | pass→pass | 19,482 | 17,397 | -11% | 1 | 1 | 0% | 2,616 | 2,020 | -23% | 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 +9 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.