Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when optimising Largest Contentful Paint (LCP), reducing render-blocking resource contention, or fine-tuning resource loading order in the critical rendering path.
.claude/skills/thedaviddias-fetchpriority-attribute/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | 37% | 0% |
| case-12 | ✓→✓ | ＝ Same ✓ | 55% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -19% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 36% | 0% |
Browsers use internal heuristics to guess resource priority, but they cannot know which image is your LCP candidate. Adding fetchpriority="high" to the LCP image has been shown to reduce LCP by 5–30 % in real-world tests, directly improving Core Web Vitals scores and user-perceived load speed. The attribute costs nothing to add and is the lowest-effort high-impact performance optimisation available today.
Check whether the LCP image or hero element has fetchpriority="high" and whether any high-priority preload links are missing the attribute.
Add fetchpriority="high" to the LCP image element and any preloaded critical resources, and fetchpriority="low" to non-critical below-the-fold images.
Explain how the browser's resource priority queue works and how fetchpriority hints change loading order to improve LCP.
Review the HTML for images above the fold. Flag the LCP candidate if it lacks fetchpriority="high", and flag any carousel or below-fold images that lack fetchpriority="low" (they compete with critical resources).
For full implementation details, code examples, and framework-specific guidance, see references/rule.md.
Rule page: https://frontendchecklist.io/en/rules/performance/fetchpriority-attribute
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | pass→pass | 4,567 | 3,986 | -13% | 1 | 1 | 0% | 922 | 1,262 | +37% | 0 | 0 | — |
case-12 | pass→pass | 2,823 | 2,249 | -20% | 1 | 1 | 0% | 482 | 745 | +55% | 0 | 0 | — |
case-01 | fail→fail | 5,530 | 3,114 | -44% | 1 | 1 | 0% | 1,073 | 991 | -8% | 0 | 0 | — |
case-02 | pass→pass | 9,404 | 6,025 | -36% | 1 | 1 | 0% | 1,891 | 1,537 | -19% | 0 | 0 | — |
case-03 | pass→pass | 3,343 | 3,892 | +16% | 1 | 1 | 0% | 600 | 817 | +36% | 0 | 0 | — |
case-04 | pass→pass | 11,552 | 7,746 | -33% | 1 | 1 | 0% | 2,023 | 1,788 | -12% | 0 | 0 | — |
case-05 | pass→pass | 5,383 | 2,216 | -59% | 1 | 1 | 0% | 1,264 | 795 | -37% | 0 | 0 | — |
case-06 | pass→pass | 8,501 | 4,132 | -51% | 1 | 1 | 0% | 1,754 | 1,193 | -32% | 0 | 0 | — |
case-08 | pass→pass | 3,446 | 2,991 | -13% | 1 | 1 | 0% | 700 | 1,087 | +55% | 0 | 0 | — |
case-09 | pass→pass | 19,732 | 13,176 | -33% | 1 | 1 | 0% | 3,064 | 2,715 | -11% | 0 | 0 | — |
case-10 | pass→pass | 11,341 | 5,535 | -51% | 1 | 1 | 0% | 2,100 | 1,390 | -34% | 0 | 0 | — |
case-11 | pass→pass | 6,987 | 4,254 | -39% | 1 | 1 | 0% | 1,617 | 1,481 | -8% | 0 | 0 | — |
case-13 | pass→pass | 8,117 | 5,892 | -27% | 1 | 1 | 0% | 1,555 | 1,405 | -10% | 0 | 0 | — |
case-14 | fail→pass | 16,607 | 8,155 | -51% | 1 | 1 | 0% | 2,956 | 2,174 | -26% | 0 | 0 | — |
case-15 | pass→pass | 12,212 | 8,838 | -28% | 1 | 1 | 0% | 2,172 | 2,011 | -7% | 0 | 0 | — |
case-16 | pass→pass | 7,316 | 5,623 | -23% | 1 | 1 | 0% | 1,677 | 1,417 | -16% | 0 | 0 | — |
case-17 | pass→pass | 5,264 | 3,767 | -28% | 1 | 1 | 0% | 928 | 1,078 | +16% | 0 | 0 | — |
case-18 | pass→pass | 3,169 | 3,131 | -1% | 1 | 1 | 0% | 736 | 922 | +25% | 0 | 0 | — |
case-19 | pass→pass | 13,075 | 12,068 | -8% | 1 | 1 | 0% | 2,307 | 2,580 | +12% | 0 | 0 | — |
case-20 | pass→pass | 9,858 | 5,878 | -40% | 1 | 1 | 0% | 1,870 | 1,567 | -16% | 0 | 0 | — |
case-21 | pass→pass | 14,638 | 10,994 | -25% | 1 | 1 | 0% | 2,470 | 2,238 | -9% | 0 | 0 | — |
case-22 | pass→pass | 20,104 | 15,717 | -22% | 1 | 1 | 0% | 2,447 | 2,967 | +21% | 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.