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 image sprites where appropriate. Check encoded size, rendered size, loading strategy, and above-the-fold impact together.
.claude/skills/thedaviddias-sprite-generation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-21 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 15% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -25% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -2% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -8% | 0% |
Each image is an HTTP request—combining many small images into one sprite reduces requests, though modern HTTP/2 and SVG icons have reduced this need.
Check if small images and icons are combined into sprites to reduce HTTP requests.
Create CSS sprites for icons and small images, or migrate to SVG icons.
Explain how sprites reduce HTTP requests but consider modern alternatives like SVG icons.
Review image assets, markup, and delivery configuration related to Use image sprites where appropriate. 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/sprite-generation
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-21 | fail→pass | 16,865 | 13,329 | -21% | 1 | 1 | 0% | 2,688 | 2,544 | -5% | 0 | 0 | — |
case-01 | pass→pass | 8,106 | 7,852 | -3% | 1 | 1 | 0% | 1,536 | 1,767 | +15% | 0 | 0 | — |
case-02 | pass→pass | 17,890 | 11,077 | -38% | 1 | 1 | 0% | 3,473 | 2,612 | -25% | 0 | 0 | — |
case-03 | pass→pass | 16,383 | 14,208 | -13% | 1 | 1 | 0% | 2,567 | 2,523 | -2% | 0 | 0 | — |
case-04 | pass→pass | 11,955 | 9,211 | -23% | 1 | 1 | 0% | 2,109 | 1,941 | -8% | 0 | 0 | — |
case-05 | pass→pass | 13,779 | 12,430 | -10% | 1 | 1 | 0% | 2,610 | 2,273 | -13% | 0 | 0 | — |
case-06 | pass→pass | 10,701 | 10,271 | -4% | 1 | 1 | 0% | 1,985 | 2,219 | +12% | 0 | 0 | — |
case-07 | pass→pass | 15,900 | 11,401 | -28% | 1 | 1 | 0% | 2,393 | 2,132 | -11% | 0 | 0 | — |
case-08 | pass→pass | 13,477 | 8,805 | -35% | 1 | 1 | 0% | 2,319 | 1,942 | -16% | 0 | 0 | — |
case-09 | pass→pass | 4,899 | 4,919 | +0% | 1 | 1 | 0% | 944 | 1,241 | +31% | 0 | 0 | — |
case-10 | pass→pass | 16,056 | 13,732 | -14% | 1 | 1 | 0% | 2,565 | 2,551 | -1% | 0 | 0 | — |
case-11 | pass→pass | 10,773 | 10,042 | -7% | 1 | 1 | 0% | 1,811 | 2,063 | +14% | 0 | 0 | — |
case-12 | pass→pass | 14,696 | 11,375 | -23% | 1 | 1 | 0% | 2,493 | 2,181 | -13% | 0 | 0 | — |
case-13 | pass→pass | 11,230 | 11,035 | -2% | 1 | 1 | 0% | 2,007 | 2,192 | +9% | 0 | 0 | — |
case-14 | pass→pass | 8,079 | 6,860 | -15% | 1 | 1 | 0% | 1,293 | 1,371 | +6% | 0 | 0 | — |
case-15 | fail→fail | 13,194 | 11,487 | -13% | 1 | 1 | 0% | 2,267 | 2,177 | -4% | 0 | 0 | — |
case-16 | pass→pass | 9,003 | 6,846 | -24% | 1 | 1 | 0% | 1,643 | 1,516 | -8% | 0 | 0 | — |
case-17 | pass→pass | 9,897 | 7,022 | -29% | 1 | 1 | 0% | 1,694 | 1,516 | -11% | 0 | 0 | — |
case-18 | pass→pass | 10,936 | 8,749 | -20% | 1 | 1 | 0% | 1,724 | 1,607 | -7% | 0 | 0 | — |
case-19 | pass→pass | 9,528 | 6,370 | -33% | 1 | 1 | 0% | 1,750 | 1,293 | -26% | 0 | 0 | — |
case-20 | pass→pass | 3,755 | 3,331 | -11% | 1 | 1 | 0% | 584 | 839 | +44% | 0 | 0 | — |
case-22 | pass→pass | 11,832 | 8,732 | -26% | 1 | 1 | 0% | 2,048 | 1,719 | -16% | 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 +5 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.