Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →按「丹心（码骨丹心）」调性写作——后端/架构/数据库方向，面试场景切入的金字塔结构。当用户要写后端面试文、Redis/Spring/MySQL 原理拆解、分布式/并发/微服务、技术选型对比、sa-token 等主题时使用。不适合纯 AI/大模型（→wei-ailang）、生活省钱/职场非技术（→wei-mingdeng）。
.claude/skills/helloworldtang-wei-danxin/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -9% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 22% | 0% |
聊后端、讲架构、拆数据库，用大白话讲透技术本质。核心读者是初中级开发者，聚焦面试。
wc -m 落在区间\\python / \\\bash）、前后空行../references/writing-standards.md 禁用词表）> 引用块）../references/writing-standards.md../references/dedup-rules.mdwei-ailang；生活省钱/职场非技术 → wei-mingdeng| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 23,531 | 27,277 | +16% | 1 | 1 | 0% | 4,128 | 5,155 | +25% | 0 | 0 | — |
case-02 | fail→fail | 23,428 | 21,205 | -9% | 1 | 1 | 0% | 3,814 | 4,277 | +12% | 0 | 0 | — |
case-03 | fail→fail | 22,317 | 28,145 | +26% | 1 | 1 | 0% | 3,378 | 5,203 | +54% | 0 | 0 | — |
case-04 | fail→pass | 32,379 | 21,789 | -33% | 1 | 1 | 0% | 4,547 | 4,913 | +8% | 0 | 0 | — |
case-05 | pass→pass | 18,756 | 22,310 | +19% | 1 | 1 | 0% | 3,177 | 4,535 | +43% | 0 | 0 | — |
case-06 | fail→pass | 23,024 | 23,445 | +2% | 1 | 1 | 0% | 4,178 | 4,908 | +17% | 0 | 0 | — |
case-07 | fail→pass | 26,725 | 20,713 | -22% | 1 | 1 | 0% | 4,309 | 3,928 | -9% | 0 | 0 | — |
case-08 | fail→pass | 22,887 | 22,588 | -1% | 1 | 1 | 0% | 3,566 | 4,342 | +22% | 0 | 0 | — |
case-09 | fail→pass | 23,499 | 25,615 | +9% | 1 | 1 | 0% | 3,518 | 4,813 | +37% | 0 | 0 | — |
case-10 | fail→pass | 26,757 | 36,237 | +35% | 1 | 1 | 0% | 4,274 | 6,519 | +53% | 0 | 0 | — |
case-11 | fail→fail | 22,607 | 31,516 | +39% | 1 | 1 | 0% | 3,876 | 6,045 | +56% | 0 | 0 | — |
case-12 | fail→fail | 26,990 | 18,833 | -30% | 1 | 1 | 0% | 4,587 | 4,296 | -6% | 0 | 0 | — |
case-13 | fail→fail | 18,317 | 29,670 | +62% | 1 | 1 | 0% | 3,101 | 5,342 | +72% | 0 | 0 | — |
case-14 | fail→fail | 17,900 | 19,366 | +8% | 1 | 1 | 0% | 3,201 | 3,645 | +14% | 0 | 0 | — |
case-15 | fail→fail | 22,242 | 26,142 | +18% | 1 | 1 | 0% | 3,914 | 5,052 | +29% | 0 | 0 | — |
case-16 | fail→pass | 28,796 | 25,086 | -13% | 1 | 1 | 0% | 4,882 | 4,917 | +1% | 0 | 0 | — |
case-17 | fail→fail | 22,689 | 29,296 | +29% | 1 | 1 | 0% | 3,985 | 6,210 | +56% | 0 | 0 | — |
case-18 | fail→pass | 25,506 | 25,996 | +2% | 1 | 1 | 0% | 3,985 | 5,463 | +37% | 0 | 0 | — |
case-19 | fail→fail | 19,464 | 21,858 | +12% | 1 | 1 | 0% | 3,395 | 4,356 | +28% | 0 | 0 | — |
case-20 | pass→pass | 27,134 | 28,068 | +3% | 1 | 1 | 0% | 4,295 | 5,483 | +28% | 0 | 0 | — |
case-21 | pass→pass | 26,341 | 28,467 | +8% | 1 | 1 | 0% | 3,707 | 4,835 | +30% | 0 | 0 | — |
case-22 | pass→pass | 19,499 | 26,778 | +37% | 1 | 1 | 0% | 3,724 | 4,995 | +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 +41 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/3/2026 | +27% |
Other measured skills in the registry, with their headline benchmark lift.