Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →用于研究任务的最终呈现形式未知时。比较研究报告、学术论文、表格优先报表、决策备忘录或自定义形式，并用权威标准与真实范例支持推荐。
.claude/skills/opensensenova-sn-report-format-discovery/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -93% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -84% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -83% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -88% | 0% |
根据用户需求确定一个请求级 format 字符串。该字符串和 language 一样只在本次运行上下文中传递，不写入文件。
querylanguagereport。report、paper、table、memo、timeline、faq，但这不是封闭枚举。format 只表示最终交付形式，不承载章节、摘要、目录、结构偏好或写作规则；这些要求继续保留在原始 query 中。format_proposal.json、format.json 或任何格式 schema。只返回一行：
textformat: <一个简短的非空字符串>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 66,254 | 7,619 | -89% | 1 | 1 | 0% | 8,213 | 572 | -93% | 0 | 0 | — |
case-02 | fail→pass | 16,388 | 8,057 | -51% | 1 | 1 | 0% | 1,951 | 591 | -70% | 0 | 0 | — |
case-03 | fail→pass | 22,252 | 6,438 | -71% | 1 | 1 | 0% | 2,884 | 459 | -84% | 0 | 0 | — |
case-04 | fail→pass | 24,174 | 6,660 | -72% | 1 | 1 | 0% | 3,208 | 535 | -83% | 0 | 0 | — |
case-05 | fail→pass | 22,889 | 6,661 | -71% | 1 | 1 | 0% | 3,786 | 472 | -88% | 0 | 0 | — |
case-06 | pass→pass | 23,909 | 7,995 | -67% | 1 | 1 | 0% | 3,404 | 569 | -83% | 0 | 0 | — |
case-07 | pass→pass | 22,438 | 8,732 | -61% | 1 | 1 | 0% | 2,965 | 627 | -79% | 0 | 0 | — |
case-08 | fail→pass | 38,470 | 7,820 | -80% | 1 | 1 | 0% | 5,378 | 606 | -89% | 0 | 0 | — |
case-09 | fail→pass | 20,568 | 6,564 | -68% | 1 | 1 | 0% | 2,504 | 491 | -80% | 0 | 0 | — |
case-10 | fail→pass | 48,478 | 6,959 | -86% | 1 | 1 | 0% | 8,215 | 572 | -93% | 0 | 0 | — |
case-11 | fail→pass | 21,784 | 7,863 | -64% | 1 | 1 | 0% | 2,766 | 745 | -73% | 0 | 0 | — |
case-12 | fail→pass | 20,048 | 7,802 | -61% | 1 | 1 | 0% | 4,022 | 688 | -83% | 0 | 0 | — |
case-13 | fail→pass | 15,530 | 7,196 | -54% | 1 | 1 | 0% | 2,427 | 586 | -76% | 0 | 0 | — |
case-14 | pass→pass | 68,429 | 8,060 | -88% | 1 | 1 | 0% | 8,205 | 597 | -93% | 0 | 0 | — |
case-15 | fail→pass | 21,754 | 7,513 | -65% | 1 | 1 | 0% | 2,538 | 640 | -75% | 0 | 0 | — |
case-16 | fail→pass | 27,906 | 7,191 | -74% | 1 | 1 | 0% | 3,313 | 580 | -82% | 0 | 0 | — |
case-17 | fail→pass | 25,367 | 6,850 | -73% | 1 | 1 | 0% | 3,296 | 492 | -85% | 0 | 0 | — |
case-18 | pass→pass | 10,934 | 2,938 | -73% | 1 | 1 | 0% | 1,786 | 787 | -56% | 0 | 0 | — |
case-19 | pass→pass | 23,960 | 2,205 | -91% | 1 | 1 | 0% | 4,194 | 581 | -86% | 0 | 0 | — |
case-20 | pass→fail | 8,007 | 8,252 | +3% | 1 | 1 | 0% | 1,719 | 571 | -67% | 0 | 0 | — |
case-21 | pass→fail | 13,556 | 4,323 | -68% | 1 | 1 | 0% | 2,209 | 1,016 | -54% | 0 | 0 | — |
case-22 | pass→fail | 30,241 | 1,982 | -93% | 1 | 1 | 0% | 3,975 | 506 | -87% | 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 +50 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are 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 | 8/4/2026 | +59% |
Other measured skills in the registry, with their headline benchmark lift.