Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when targeting 《经济科学》(Economic Science) or deciding whether a Chinese social-science/econ/management manuscript fits this journal. Applies the journal's fit, framing, method, house-style, official-submission-check, and desk-reject heuristics.
.claude/skills/brycewang-stanford-economic-science/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-13 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 1% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 12% | 0% |
北京大学主办的综合经济学期刊，官方稿约强调学术导向、科学问题和文献价值。
这个 skill 是定位 / 选刊 / 改写框架工具，不替代该刊最新官方投稿须知。正式投稿前必须重新核对官网、采编系统或编辑部发布的最新模板。
economic-research（《经济研究》） / economic-review-cn（《经济评论》） / economic-theory-and-business-management（《经济理论与经济管理》） / economist-cn（《经济学家》）。若这些刊物的读者对象更贴近，不要因为名称、地区或变量相似而强行投本刊。../../resources/source-basis.md 和 ../../resources/official-source-map.md，从其中的官方来源锚点或同一主办/出版体系入口开始核验，并在回答中说明核验日期。方法更前沿可投 CEQ；理论贡献和中国问题更厚可投 economic-research。
text【匹配度】高 / 中 / 低（一句话理由） 【目标期刊】《经济科学》 【选题标签】<最贴近的 2-3 个主题> 【方法证据】<当前方法是否够本刊标准> 【最大风险】<最可能导致退稿的一点> 【需核验官方要求】<投稿系统/匿名/摘要/图表/数据等> 【改投建议】<若不匹配，给出更合适期刊>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-13 | fail→pass | 25,988 | 22,159 | -15% | 1 | 1 | 0% | 2,584 | 3,468 | +34% | 0 | 0 | — |
case-14 | fail→fail | 26,026 | 26,548 | +2% | 1 | 1 | 0% | 2,841 | 3,824 | +35% | 0 | 0 | — |
case-15 | fail→fail | 33,247 | 19,294 | -42% | 1 | 1 | 0% | 3,453 | 3,003 | -13% | 0 | 0 | — |
case-21 | pass→pass | 8,546 | 13,245 | +55% | 1 | 1 | 0% | 1,401 | 2,413 | +72% | 0 | 0 | — |
case-01 | fail→pass | 25,323 | 24,255 | -4% | 1 | 1 | 0% | 3,000 | 3,442 | +15% | 0 | 0 | — |
case-02 | fail→fail | 35,061 | 21,494 | -39% | 1 | 1 | 0% | 3,655 | 3,402 | -7% | 0 | 0 | — |
case-03 | fail→fail | 23,353 | 14,959 | -36% | 1 | 1 | 0% | 2,621 | 2,942 | +12% | 0 | 0 | — |
case-04 | fail→pass | 28,190 | 24,621 | -13% | 1 | 1 | 0% | 3,446 | 3,371 | -2% | 0 | 0 | — |
case-05 | fail→pass | 30,172 | 22,096 | -27% | 1 | 1 | 0% | 3,121 | 3,151 | +1% | 0 | 0 | — |
case-06 | pass→pass | 21,375 | 22,523 | +5% | 1 | 1 | 0% | 2,195 | 3,834 | +75% | 0 | 0 | — |
case-07 | pass→pass | 29,784 | 27,163 | -9% | 1 | 1 | 0% | 3,024 | 3,855 | +27% | 0 | 0 | — |
case-08 | fail→fail | 22,168 | 32,664 | +47% | 1 | 1 | 0% | 3,168 | 4,171 | +32% | 0 | 0 | — |
case-09 | fail→fail | 23,559 | 24,924 | +6% | 1 | 1 | 0% | 2,654 | 3,357 | +26% | 0 | 0 | — |
case-10 | fail→fail | 33,219 | 25,277 | -24% | 1 | 1 | 0% | 3,488 | 3,587 | +3% | 0 | 0 | — |
case-11 | fail→pass | 25,980 | 21,650 | -17% | 1 | 1 | 0% | 2,871 | 3,212 | +12% | 0 | 0 | — |
case-12 | fail→pass | 32,811 | 32,823 | +0% | 1 | 1 | 0% | 3,498 | 4,403 | +26% | 0 | 0 | — |
case-16 | fail→fail | 24,133 | 16,213 | -33% | 1 | 1 | 0% | 2,993 | 3,475 | +16% | 0 | 0 | — |
case-17 | fail→pass | 39,049 | 20,568 | -47% | 1 | 1 | 0% | 2,676 | 3,197 | +19% | 0 | 0 | — |
case-18 | fail→pass | 26,134 | 24,618 | -6% | 1 | 1 | 0% | 2,852 | 3,566 | +25% | 0 | 0 | — |
case-19 | fail→pass | 27,377 | 26,428 | -3% | 1 | 1 | 0% | 2,786 | 3,692 | +33% | 0 | 0 | — |
case-20 | pass→pass | 19,302 | 22,820 | +18% | 1 | 1 | 0% | 1,970 | 3,198 | +62% | 0 | 0 | — |
case-22 | pass→pass | 20,934 | 23,915 | +14% | 1 | 1 | 0% | 2,835 | 4,328 | +53% | 0 | 0 | — |
case-23 | pass→pass | 30,522 | 15,054 | -51% | 1 | 1 | 0% | 4,730 | 3,305 | -30% | 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. 23 cases were attempted. The headline lift of +39 percentage points is the difference between those two pass rates over the 23 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.