Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →检查进行中的假期，获取截止日期预警和需要做出的决策。仅呈现 需要采取行动的假期并说明原因——不是状态面板。建议每周运行， 或每当律师需要知道哪些假期有即将到来的审批、证明或到期截止时使用。
.claude/skills/thomasmoreai-leave-tracker-zhou210712/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -10% | 0% |
检查所有有法定硬性截止日期的进行中假期，仅呈现需要决策或行动的项目。不是状态面板——告诉你需要做什么及为什么。
~/.claude/plugins/config/claude-for-legal/employment-legal/CLAUDE.md → 管辖地表和假期管理部分。~/.claude/plugins/config/claude-for-legal/employment-legal/leave-register.yaml 不存在或无数据，提示律师上传假期电子表格或使用 /employment-legal:log-leave 添加条目。| 假期类型 | 关键截止日期 | 法律依据 | |---|---|---| | 医疗期 | 医疗期满日（按工作年限3-24个月） | 企业职工患病或非因工负伤医疗期规定 [法条原文] | | 产假 | 产假期满日（98天基础 + 省/直辖市奖励假） | 女职工劳动保护特别规定 [法条原文] | | 年休假 | 年末未休结转截止（一般不跨年结转） | 职工带薪年休假条例第5条 [法条原文] | | 工伤假 | 停工留薪期满日（一般不超过12个月） | 工伤保险条例第33条 [法条原文] | | 婚假 | 各省/直辖市人口与计划生育条例规定 | 省/直辖市规定 |
/employment-legal:leave-tracker建议每周运行——设置周一上午提醒调用 /employment-legal:leave-tracker。自动排期需要单独的集成（日历提醒、定时任务等）；Claude Code 不自行排期。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-14 | fail→pass | 11,995 | 1,946 | -84% | 1 | 1 | 0% | 2,004 | 774 | -61% | 0 | 0 | — |
case-01 | pass→pass | 8,201 | 11,147 | +36% | 1 | 1 | 0% | 1,318 | 2,114 | +60% | 0 | 0 | — |
case-02 | pass→pass | 11,055 | 15,139 | +37% | 1 | 1 | 0% | 1,809 | 3,004 | +66% | 0 | 0 | — |
case-03 | fail→fail | 19,168 | 24,283 | +27% | 1 | 1 | 0% | 3,034 | 3,809 | +26% | 0 | 0 | — |
case-04 | pass→pass | 10,679 | 5,165 | -52% | 1 | 1 | 0% | 1,537 | 1,219 | -21% | 0 | 0 | — |
case-05 | pass→pass | 8,870 | 2,959 | -67% | 1 | 1 | 0% | 1,603 | 979 | -39% | 0 | 0 | — |
case-06 | pass→pass | 7,249 | 2,575 | -64% | 1 | 1 | 0% | 1,064 | 875 | -18% | 0 | 0 | — |
case-07 | pass→pass | 6,764 | 3,668 | -46% | 1 | 1 | 0% | 1,135 | 1,072 | -6% | 0 | 0 | — |
case-08 | pass→pass | 14,392 | 14,998 | +4% | 1 | 1 | 0% | 2,321 | 2,969 | +28% | 0 | 0 | — |
case-09 | pass→pass | 16,041 | 10,127 | -37% | 1 | 1 | 0% | 2,584 | 2,345 | -9% | 0 | 0 | — |
case-10 | pass→pass | 5,125 | 3,718 | -27% | 1 | 1 | 0% | 784 | 1,000 | +28% | 0 | 0 | — |
case-11 | pass→pass | 13,854 | 34,647 | +150% | 1 | 1 | 0% | 2,415 | 2,194 | -9% | 0 | 0 | — |
case-12 | pass→pass | 19,261 | 10,380 | -46% | 1 | 1 | 0% | 2,468 | 2,175 | -12% | 0 | 0 | — |
case-13 | fail→pass | 10,759 | 1,975 | -82% | 1 | 1 | 0% | 1,700 | 694 | -59% | 0 | 0 | — |
case-15 | fail→pass | 13,374 | 8,227 | -38% | 1 | 1 | 0% | 2,258 | 1,806 | -20% | 0 | 0 | — |
case-16 | fail→pass | 13,230 | 12,790 | -3% | 1 | 1 | 0% | 2,212 | 2,608 | +18% | 0 | 0 | — |
case-17 | fail→pass | 16,664 | 9,299 | -44% | 1 | 1 | 0% | 2,182 | 1,962 | -10% | 0 | 0 | — |
case-18 | fail→pass | 15,576 | 6,390 | -59% | 1 | 1 | 0% | 2,171 | 1,436 | -34% | 0 | 0 | — |
case-19 | pass→pass | 16,727 | 15,080 | -10% | 1 | 1 | 0% | 2,797 | 2,757 | -1% | 0 | 0 | — |
case-20 | pass→pass | 14,871 | 8,605 | -42% | 1 | 1 | 0% | 1,303 | 2,038 | +56% | 0 | 0 | — |
case-21 | pass→pass | 4,029 | 6,092 | +51% | 1 | 1 | 0% | 711 | 1,609 | +126% | 0 | 0 | — |
case-22 | fail→pass | 16,875 | 59,403 | +252% | 1 | 1 | 0% | 2,322 | 3,276 | +41% | 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 +32 percentage points is the difference between those two pass rates over the 22 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.