Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when reviewing templates, rendered HTML, or shared components related to Add resource hints (preload, prefetch, dns-prefetch). Validate the final browser-facing markup, not just the source framework abstraction.
.claude/skills/thedaviddias-html-resource-hints/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 41% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -12% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -3% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 24% | 0% |
The browser's HTML parser discovers resources sequentially — a font referenced in a CSS file won't be fetched until the CSS is parsed. Resource hints break this dependency chain, allowing the browser to start fetching critical resources in parallel with parsing. A single preload directive for a critical font can eliminate a 100–300ms Flash of Invisible Text.
Look at this HTML file for late-discovered critical resources — fonts referenced in CSS, hero images, and key scripts. Identify where preload or preconnect hints would help.
Add appropriate preload hints for critical fonts, images, and scripts. Add preconnect for third-party domains used early in the page.
Explain each resource hint type (preload, prefetch, preconnect, dns-prefetch), when to use each, and the consequences of misusing them.
Review templates, server-rendered HTML, and shared components that output markup related to Add resource hints (preload, prefetch, dns-prefetch). Flag exact elements, attributes, and routes where the rendered HTML violates the rule.
For full implementation details, code examples, and framework-specific guidance, see references/rule.md.
Rule page: https://frontendchecklist.io/en/rules/html/html-resource-hints
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 12,619 | 11,791 | -7% | 1 | 1 | 0% | 2,950 | 2,559 | -13% | 0 | 0 | — |
case-02 | fail→fail | 11,644 | 11,941 | +3% | 1 | 1 | 0% | 2,388 | 2,910 | +22% | 0 | 0 | — |
case-03 | pass→pass | 6,000 | 6,764 | +13% | 1 | 1 | 0% | 1,232 | 1,737 | +41% | 0 | 0 | — |
case-04 | pass→pass | 10,556 | 7,706 | -27% | 1 | 1 | 0% | 2,066 | 1,812 | -12% | 0 | 0 | — |
case-05 | pass→pass | 5,111 | 2,855 | -44% | 1 | 1 | 0% | 902 | 874 | -3% | 0 | 0 | — |
case-06 | pass→pass | 5,485 | 5,705 | +4% | 1 | 1 | 0% | 1,231 | 1,526 | +24% | 0 | 0 | — |
case-07 | pass→pass | 10,604 | 9,318 | -12% | 1 | 1 | 0% | 2,044 | 2,140 | +5% | 0 | 0 | — |
case-08 | pass→pass | 6,976 | 6,780 | -3% | 1 | 1 | 0% | 1,671 | 1,820 | +9% | 0 | 0 | — |
case-09 | pass→pass | 12,739 | 8,600 | -32% | 1 | 1 | 0% | 2,396 | 2,097 | -12% | 0 | 0 | — |
case-10 | pass→pass | 2,696 | 2,389 | -11% | 1 | 1 | 0% | 578 | 823 | +42% | 0 | 0 | — |
case-11 | pass→pass | 3,477 | 2,893 | -17% | 1 | 1 | 0% | 688 | 1,020 | +48% | 0 | 0 | — |
case-12 | pass→pass | 7,657 | 6,384 | -17% | 1 | 1 | 0% | 1,551 | 1,649 | +6% | 0 | 0 | — |
case-13 | pass→pass | 6,787 | 4,626 | -32% | 1 | 1 | 0% | 1,368 | 1,248 | -9% | 0 | 0 | — |
case-14 | pass→pass | 4,799 | 3,142 | -35% | 1 | 1 | 0% | 1,022 | 1,102 | +8% | 0 | 0 | — |
case-15 | pass→pass | 10,908 | 6,690 | -39% | 1 | 1 | 0% | 1,793 | 1,643 | -8% | 0 | 0 | — |
case-16 | pass→pass | 8,569 | 8,675 | +1% | 1 | 1 | 0% | 1,594 | 1,882 | +18% | 0 | 0 | — |
case-17 | pass→pass | 7,887 | 8,662 | +10% | 1 | 1 | 0% | 1,760 | 2,075 | +18% | 0 | 0 | — |
case-18 | pass→pass | 10,723 | 10,360 | -3% | 1 | 1 | 0% | 2,240 | 2,598 | +16% | 0 | 0 | — |
case-19 | pass→pass | 12,686 | 8,989 | -29% | 1 | 1 | 0% | 2,331 | 1,986 | -15% | 0 | 0 | — |
case-20 | pass→pass | 8,921 | 5,922 | -34% | 1 | 1 | 0% | 1,741 | 1,537 | -12% | 0 | 0 | — |
case-21 | pass→pass | 7,420 | 5,074 | -32% | 1 | 1 | 0% | 1,576 | 1,569 | -0% | 0 | 0 | — |
case-22 | pass→pass | 4,114 | 4,669 | +13% | 1 | 1 | 0% | 963 | 1,294 | +34% | 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.