Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Google Workspace API操作、Sheets自動化、Gmail統合、およびドキュメント管理。
.claude/skills/affaan-m-google-workspace-ops/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 1% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 27% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | -5% | 0% |
此技能用于将共享文档、电子表格和演示文稿作为工作系统进行操作，而不仅仅是孤立地编辑单个文件。
使用 Google Drive 作为入口，然后切换到合适的专业工具：
不要仅凭文件名猜测结构。先检查。
从 Drive 搜索界面开始，定位：
如果多个文档看起来相似，请通过标题、所有者、修改时间或文件夹进行确认。
在进行更改之前：
选择能够安全完成工作的最小工具。
如果请求的工作涉及视觉或布局调整，请通过检查和验证进行迭代，而不是进行一次性的盲目更新。
当文件是更大工作流程的一部分时，还需指出：
使用：
text资产 - 文件名 - 类型 - 为何选择此文件 当前状态 - 结构摘要 - 关键问题或阻碍 操作 - 已执行或建议的编辑 后续事项 - 归档 / 合并 / 重复清理 / 下一个待更新文件
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 11,274 | 6,683 | -41% | 1 | 1 | 0% | 1,694 | 1,602 | -5% | 0 | 0 | — |
case-01 | pass→pass | 15,807 | 10,768 | -32% | 1 | 1 | 0% | 2,304 | 2,336 | +1% | 0 | 0 | — |
case-02 | pass→pass | 18,878 | 12,687 | -33% | 1 | 1 | 0% | 3,000 | 2,809 | -6% | 0 | 0 | — |
case-03 | pass→pass | 13,600 | 13,957 | +3% | 1 | 1 | 0% | 2,415 | 2,929 | +21% | 0 | 0 | — |
case-04 | pass→fail | 10,786 | 12,354 | +15% | 1 | 1 | 0% | 2,153 | 2,734 | +27% | 0 | 0 | — |
case-05 | pass→pass | 11,745 | 12,637 | +8% | 1 | 1 | 0% | 2,159 | 2,840 | +32% | 0 | 0 | — |
case-07 | pass→pass | 18,027 | 6,043 | -66% | 1 | 1 | 0% | 1,772 | 1,271 | -28% | 0 | 0 | — |
case-08 | fail→pass | 14,231 | 8,515 | -40% | 1 | 1 | 0% | 2,137 | 1,777 | -17% | 0 | 0 | — |
case-09 | pass→pass | 12,680 | 15,645 | +23% | 1 | 1 | 0% | 1,791 | 2,787 | +56% | 0 | 0 | — |
case-10 | pass→pass | 9,349 | 8,490 | -9% | 1 | 1 | 0% | 1,459 | 1,968 | +35% | 0 | 0 | — |
case-11 | fail→pass | 8,210 | 4,037 | -51% | 1 | 1 | 0% | 1,152 | 1,141 | -1% | 0 | 0 | — |
case-12 | pass→pass | 11,556 | 5,350 | -54% | 1 | 1 | 0% | 1,626 | 1,293 | -20% | 0 | 0 | — |
case-13 | pass→pass | 7,310 | 2,846 | -61% | 1 | 1 | 0% | 1,099 | 1,000 | -9% | 0 | 0 | — |
case-14 | pass→pass | 3,113 | 3,320 | +7% | 1 | 1 | 0% | 456 | 1,035 | +127% | 0 | 0 | — |
case-15 | pass→pass | 6,744 | 3,876 | -43% | 1 | 1 | 0% | 1,131 | 1,160 | +3% | 0 | 0 | — |
case-16 | pass→pass | 5,748 | 5,341 | -7% | 1 | 1 | 0% | 911 | 1,306 | +43% | 0 | 0 | — |
case-17 | pass→pass | 10,583 | 3,148 | -70% | 1 | 1 | 0% | 1,553 | 1,064 | -31% | 0 | 0 | — |
case-18 | pass→pass | 10,164 | 6,983 | -31% | 1 | 1 | 0% | 1,633 | 1,620 | -1% | 0 | 0 | — |
case-19 | pass→pass | 9,175 | 9,090 | -1% | 1 | 1 | 0% | 1,461 | 1,872 | +28% | 0 | 0 | — |
case-20 | pass→pass | 10,641 | 8,736 | -18% | 1 | 1 | 0% | 1,690 | 1,933 | +14% | 0 | 0 | — |
case-21 | pass→pass | 11,328 | 9,613 | -15% | 1 | 1 | 0% | 1,655 | 2,014 | +22% | 0 | 0 | — |
case-22 | fail→pass | 10,969 | 9,677 | -12% | 1 | 1 | 0% | 1,890 | 1,916 | +1% | 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 +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.
Other measured skills in the registry, with their headline benchmark lift.