Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →为中文毕业论文补强并核验引用：找出必须有引文支撑的句子，扩展候选文献，检查引用与论断是否匹配，并回写参考文献与正文引用。 **Trigger**: 引用补强, citation enhance, 文献补充, 引用核验, 毕业论文参考文献检查. **Use when**: 正文已有一定稳定度，需要系统补足背景、定义、对照工作和关键结论的引用支撑，并检查是否存在误引或漏引。 **Skip if**: 还没有稳定正文，或当前仅在做早期结构重构。 **Network**: optional. **Guardrail**: 不为了堆引用而堆引用；引用必须与论断匹配；先补对，再补多。
.claude/skills/willoscar-thesis-citation-enhance-review/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -78% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -74% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -49% | 0% |
references/*.bibAlways read:
references/overview.mdreferences/citation_priority.mdreferences/review_rules.mdMachine-readable contract:
assets/citation_enhance_contract.jsonbib 与正文引用| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,549 | 6,603 | +1% | 1 | 1 | 0% | 1,036 | 1,185 | +14% | 0 | 0 | — |
case-02 | fail→pass | 8,730 | 1,633 | -81% | 1 | 1 | 0% | 1,213 | 375 | -69% | 0 | 0 | — |
case-03 | fail→pass | 30,579 | 1,892 | -94% | 1 | 1 | 0% | 1,695 | 370 | -78% | 0 | 0 | — |
case-04 | fail→pass | 6,232 | 1,859 | -70% | 1 | 1 | 0% | 890 | 388 | -56% | 0 | 0 | — |
case-05 | fail→pass | 21,807 | 1,726 | -92% | 1 | 1 | 0% | 1,299 | 339 | -74% | 0 | 0 | — |
case-06 | fail→fail | 4,070 | 2,667 | -34% | 1 | 1 | 0% | 553 | 481 | -13% | 0 | 0 | — |
case-07 | fail→fail | 7,419 | 2,166 | -71% | 1 | 1 | 0% | 1,013 | 450 | -56% | 0 | 0 | — |
case-08 | fail→pass | 7,034 | 3,017 | -57% | 1 | 1 | 0% | 1,108 | 569 | -49% | 0 | 0 | — |
case-09 | fail→fail | 13,451 | 4,948 | -63% | 1 | 1 | 0% | 1,972 | 903 | -54% | 0 | 0 | — |
case-10 | fail→fail | 13,304 | 8,322 | -37% | 1 | 1 | 0% | 1,925 | 1,265 | -34% | 0 | 0 | — |
case-11 | fail→fail | 6,450 | 3,219 | -50% | 1 | 1 | 0% | 993 | 643 | -35% | 0 | 0 | — |
case-12 | fail→fail | 8,622 | 2,141 | -75% | 1 | 1 | 0% | 1,191 | 421 | -65% | 0 | 0 | — |
case-13 | pass→pass | 13,828 | 2,164 | -84% | 1 | 1 | 0% | 1,926 | 416 | -78% | 0 | 0 | — |
case-14 | fail→pass | 13,706 | 2,622 | -81% | 1 | 1 | 0% | 2,040 | 543 | -73% | 0 | 0 | — |
case-15 | pass→pass | 10,023 | 4,282 | -57% | 1 | 1 | 0% | 1,592 | 860 | -46% | 0 | 0 | — |
case-16 | pass→pass | 11,755 | 4,831 | -59% | 1 | 1 | 0% | 1,870 | 852 | -54% | 0 | 0 | — |
case-17 | fail→pass | 8,859 | 4,570 | -48% | 1 | 1 | 0% | 1,287 | 802 | -38% | 0 | 0 | — |
case-18 | pass→pass | 10,164 | 4,357 | -57% | 1 | 1 | 0% | 1,375 | 773 | -44% | 0 | 0 | — |
case-19 | pass→pass | 9,999 | 8,581 | -14% | 1 | 1 | 0% | 1,518 | 1,373 | -10% | 0 | 0 | — |
case-20 | pass→pass | 10,182 | 8,308 | -18% | 1 | 1 | 0% | 1,708 | 1,530 | -10% | 0 | 0 | — |
case-21 | pass→pass | 5,528 | 5,771 | +4% | 1 | 1 | 0% | 912 | 1,091 | +20% | 0 | 0 | — |
case-22 | pass→pass | 11,941 | 12,077 | +1% | 1 | 1 | 0% | 2,105 | 2,422 | +15% | 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, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +32 percentage points is the difference between those two pass rates over the 21 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.