Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →向进行中的调查添加数据——文件、访谈记录或观察意见。 按已记录的筛选标准批量处理，浮现重要事项，记录所有已审查内容 以供覆盖验证。当新的证据、访谈记录或文件材料进入进行中的调查时使用。
.claude/skills/thomasmoreai-investigation-add-zhou210712/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -59% | 0% |
向进行中的调查日志添加数据。使用已记录的筛选标准处理文件批次，浮现重要事项，记录所有已审查内容以供覆盖验证。
~/.claude/plugins/config/claude-for-legal/employment-legal/CLAUDE.md。添加数据时，按以下维度评估：
| 维度 | 检查 | |---|---| | 相关性 | 该证据是否与待证事实相关？ | | 真实性 | 该证据是否可核验来源和真实性？ | | 合法性 | 该证据的获取方式是否合法？（《劳动争议司法解释（一）》关于证据合法性） | | 重要性 | 该证据是否会影响调查结论？ |
所有添加的数据受《律师法》第38条保密义务保护。[法条原文] 注意区分：
/employment-legal:investigation-add [调查事项名称]
[粘贴访谈记录]/employment-legal:investigation-add [调查事项名称]
[附上邮件导出文件]| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-13 | pass→pass | 13,253 | 7,924 | -40% | 1 | 1 | 0% | 2,094 | 1,672 | -20% | 0 | 0 | — |
case-01 | fail→pass | 9,563 | 7,210 | -25% | 1 | 1 | 0% | 1,405 | 1,849 | +32% | 0 | 0 | — |
case-02 | fail→fail | 20,603 | 8,327 | -60% | 1 | 1 | 0% | 1,617 | 1,750 | +8% | 0 | 0 | — |
case-03 | pass→pass | 21,849 | 15,884 | -27% | 1 | 1 | 0% | 3,041 | 2,987 | -2% | 0 | 0 | — |
case-04 | fail→pass | 20,162 | 17,395 | -14% | 1 | 1 | 0% | 3,170 | 3,181 | +0% | 0 | 0 | — |
case-05 | fail→pass | 22,653 | 13,382 | -41% | 1 | 1 | 0% | 2,620 | 2,507 | -4% | 0 | 0 | — |
case-06 | pass→pass | 13,429 | 6,631 | -51% | 1 | 1 | 0% | 1,960 | 1,501 | -23% | 0 | 0 | — |
case-07 | fail→fail | 8,230 | 3,880 | -53% | 1 | 1 | 0% | 1,483 | 944 | -36% | 0 | 0 | — |
case-08 | fail→pass | 12,536 | 3,761 | -70% | 1 | 1 | 0% | 2,281 | 873 | -62% | 0 | 0 | — |
case-09 | pass→pass | 15,281 | 13,628 | -11% | 1 | 1 | 0% | 2,039 | 2,361 | +16% | 0 | 0 | — |
case-10 | pass→pass | 12,305 | 9,294 | -24% | 1 | 1 | 0% | 1,892 | 1,841 | -3% | 0 | 0 | — |
case-11 | fail→pass | 12,919 | 1,635 | -87% | 1 | 1 | 0% | 1,647 | 677 | -59% | 0 | 0 | — |
case-12 | pass→pass | 17,847 | 7,409 | -58% | 1 | 1 | 0% | 2,688 | 1,564 | -42% | 0 | 0 | — |
case-14 | fail→pass | 13,192 | 4,241 | -68% | 1 | 1 | 0% | 2,048 | 991 | -52% | 0 | 0 | — |
case-15 | fail→pass | 14,262 | 14,973 | +5% | 1 | 1 | 0% | 2,254 | 2,942 | +31% | 0 | 0 | — |
case-16 | pass→pass | 18,351 | 11,957 | -35% | 1 | 1 | 0% | 3,034 | 2,451 | -19% | 0 | 0 | — |
case-17 | fail→pass | 14,232 | 5,671 | -60% | 1 | 1 | 0% | 2,456 | 1,230 | -50% | 0 | 0 | — |
case-18 | fail→pass | 11,915 | 5,647 | -53% | 1 | 1 | 0% | 1,782 | 1,241 | -30% | 0 | 0 | — |
case-19 | fail→pass | 16,129 | 11,562 | -28% | 1 | 1 | 0% | 2,443 | 2,319 | -5% | 0 | 0 | — |
case-20 | pass→pass | 26,371 | 24,066 | -9% | 1 | 1 | 0% | 3,776 | 3,972 | +5% | 0 | 0 | — |
case-21 | pass→pass | 23,292 | 25,459 | +9% | 1 | 1 | 0% | 3,500 | 4,561 | +30% | 0 | 0 | — |
case-22 | pass→pass | 16,396 | 15,038 | -8% | 1 | 1 | 0% | 2,748 | 2,861 | +4% | 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 +45 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.