Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Next.js 16+ and Turbopack — incremental bundling, FS caching, dev speed, and when to use Turbopack vs webpack.
.claude/skills/affaan-m-nextjs-turbopack/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -4% | 0% |
Next.js 16+ 在本地开发中默认使用 Turbopack：这是一个用 Rust 编写的增量捆绑器，能显著加快开发启动和热更新的速度。
--webpack（或 --no-turbopack，具体取决于你的 Next.js 版本；请查阅你所用版本的文档）来禁用。next build) 可能使用 Turbopack 或 webpack，这取决于 Next.js 版本；请查阅你所用版本的官方 Next.js 文档。适用场景：开发或调试 Next.js 16+ 应用，诊断开发启动或热模块替换速度慢的问题，或优化生产环境捆绑包。
next dev 默认使用 Turbopack，除非被禁用。.next 下；基本使用无需额外配置。bashnext dev next build next start
运行 next dev 以使用 Turbopack 进行本地开发。使用捆绑包分析器（参见 Next.js 文档）来优化代码分割并剔除大型依赖。尽可能优先使用 App Router 和服务器组件。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 17,217 | 18,197 | +6% | 1 | 1 | 0% | 3,229 | 3,565 | +10% | 0 | 0 | — |
case-02 | fail→fail | 18,889 | 15,346 | -19% | 1 | 1 | 0% | 3,468 | 3,482 | +0% | 0 | 0 | — |
case-03 | pass→pass | 7,147 | 3,729 | -48% | 1 | 1 | 0% | 1,279 | 1,234 | -4% | 0 | 0 | — |
case-04 | pass→pass | 4,658 | 3,226 | -31% | 1 | 1 | 0% | 906 | 1,130 | +25% | 0 | 0 | — |
case-05 | pass→pass | 4,657 | 3,070 | -34% | 1 | 1 | 0% | 715 | 1,111 | +55% | 0 | 0 | — |
case-06 | fail→pass | 8,913 | 7,413 | -17% | 1 | 1 | 0% | 1,622 | 1,797 | +11% | 0 | 0 | — |
case-07 | fail→fail | 11,815 | 10,181 | -14% | 1 | 1 | 0% | 2,289 | 2,524 | +10% | 0 | 0 | — |
case-08 | pass→pass | 11,163 | 9,262 | -17% | 1 | 1 | 0% | 1,788 | 2,135 | +19% | 0 | 0 | — |
case-09 | pass→pass | 13,940 | 9,881 | -29% | 1 | 1 | 0% | 2,377 | 2,211 | -7% | 0 | 0 | — |
case-10 | fail→pass | 14,947 | 13,524 | -10% | 1 | 1 | 0% | 2,433 | 2,917 | +20% | 0 | 0 | — |
case-11 | fail→pass | 15,368 | 5,291 | -66% | 1 | 1 | 0% | 2,462 | 1,291 | -48% | 0 | 0 | — |
case-12 | fail→pass | 8,618 | 2,857 | -67% | 1 | 1 | 0% | 1,057 | 991 | -6% | 0 | 0 | — |
case-13 | pass→pass | 18,814 | 3,386 | -82% | 1 | 1 | 0% | 3,475 | 1,085 | -69% | 0 | 0 | — |
case-14 | pass→pass | 10,987 | 6,525 | -41% | 1 | 1 | 0% | 2,067 | 1,752 | -15% | 0 | 0 | — |
case-15 | pass→pass | 6,510 | 3,308 | -49% | 1 | 1 | 0% | 1,152 | 1,038 | -10% | 0 | 0 | — |
case-16 | pass→pass | 8,594 | 6,442 | -25% | 1 | 1 | 0% | 1,493 | 1,653 | +11% | 0 | 0 | — |
case-17 | pass→pass | 8,645 | 7,814 | -10% | 1 | 1 | 0% | 1,576 | 1,919 | +22% | 0 | 0 | — |
case-18 | pass→pass | 12,032 | 2,918 | -76% | 1 | 1 | 0% | 1,765 | 1,039 | -41% | 0 | 0 | — |
case-19 | pass→pass | 11,795 | 5,889 | -50% | 1 | 1 | 0% | 2,018 | 1,499 | -26% | 0 | 0 | — |
case-20 | pass→pass | 4,940 | 4,832 | -2% | 1 | 1 | 0% | 999 | 1,467 | +47% | 0 | 0 | — |
case-21 | pass→pass | 11,368 | 11,959 | +5% | 1 | 1 | 0% | 2,248 | 2,846 | +27% | 0 | 0 | — |
case-22 | pass→pass | 4,854 | 4,158 | -14% | 1 | 1 | 0% | 874 | 1,350 | +54% | 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 +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.