Install any skill in seconds. Free to start, no credit card required.
Get Started Free →根据 Excel 数据量级自动判断处理策略,执行数值列清洗、条件过滤,并使用 openpyxl 对符合条件的单元格进行样式标记与导出。
.claude/skills/opensensenova-excel-threshold-analysis-and-styling/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 8% | 0% |
> Note: This sub-skill covers one step of the Excel analysis workflow. For the full pipeline (file reading, row counting, large-file optimization, export), see the parent workflow SKILL.md.
Step1 读取 Excel 文件中所有工作表的行数并汇总,用于评估数据规模。
pythonimport pandas as pd file_path = 'input_file.xlsx' # 读取所有 sheet 名称并统计总行数 xls = pd.ExcelFile(file_path) sheet_names = xls.sheet_names total_rows = 0 for sheet in sheet_names: # header=None 用于快速统计包含表头的总行数 df_tmp = pd.read_excel(file_path, sheet_name=sheet, header=None) rows = len(df_tmp) total_rows += rows print(f"Sheet '{sheet}': {rows} 行") print(f"\n总行数汇总: {total_rows}")
Step2 对目标数据表进行清洗,将指定列的非数值内容转换为缺失值并剔除,确保数据类型为数值型。
pythontarget_sheet = 'Sheet1' target_col = '数量' # 待处理的目标列名 header_idx = 1 # 表头所在行索引(0开始计数) df = pd.read_excel(file_path, sheet_name=target_sheet, header=header_idx) # 强制转换数值类型,无法转换的内容变为 NaN 并删除 df[target_col] = pd.to_numeric(df[target_col], errors='coerce') df_cleaned = df.dropna(subset=[target_col]) print(f"清洗完成,有效数据行数: {len(df_cleaned)}")
Step3 筛选符合特定数值条件的记录并进行统计。
pythonfilter_threshold = 10 df_filtered = df_cleaned[df_cleaned[target_col] > filter_threshold] print(f"{target_col} 大于 {filter_threshold} 的记录共有 {len(df_filtered)} 条")
Step4 使用 openpyxl 对原始文件中
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 7,463 | 11,805 | +58% | 1 | 1 | 0% | 1,699 | 2,366 | +39% | 0 | 0 | — |
case-01 | fail→pass | 14,345 | 8,885 | -38% | 1 | 1 | 0% | 2,769 | 2,394 | -14% | 0 | 0 | — |
case-02 | fail→pass | 9,077 | 9,059 | -0% | 1 | 1 | 0% | 1,905 | 2,017 | +6% | 0 | 0 | — |
case-03 | fail→pass | 16,087 | 7,706 | -52% | 1 | 1 | 0% | 3,393 | 2,279 | -33% | 0 | 0 | — |
case-04 | pass→pass | 13,736 | 8,550 | -38% | 1 | 1 | 0% | 2,641 | 2,191 | -17% | 0 | 0 | — |
case-05 | pass→pass | 9,695 | 7,173 | -26% | 1 | 1 | 0% | 1,865 | 1,842 | -1% | 0 | 0 | — |
case-06 | pass→pass | 11,868 | 10,626 | -10% | 1 | 1 | 0% | 2,563 | 2,694 | +5% | 0 | 0 | — |
case-08 | fail→pass | 8,653 | 6,920 | -20% | 1 | 1 | 0% | 1,760 | 1,893 | +8% | 0 | 0 | — |
case-09 | fail→pass | 6,049 | 11,730 | +94% | 1 | 1 | 0% | 281 | 3,255 | +1058% | 0 | 0 | — |
case-10 | fail→pass | 6,463 | 6,072 | -6% | 1 | 1 | 0% | 952 | 1,466 | +54% | 0 | 0 | — |
case-11 | fail→pass | 11,629 | 6,643 | -43% | 1 | 1 | 0% | 1,924 | 1,945 | +1% | 0 | 0 | — |
case-12 | fail→pass | 7,917 | 6,679 | -16% | 1 | 1 | 0% | 697 | 1,912 | +174% | 0 | 0 | — |
case-13 | fail→pass | 5,644 | 14,020 | +148% | 1 | 1 | 0% | 1,201 | 2,777 | +131% | 0 | 0 | — |
case-14 | fail→pass | 20,336 | 5,035 | -75% | 1 | 1 | 0% | 855 | 1,522 | +78% | 0 | 0 | — |
case-15 | fail→fail | 9,038 | 11,842 | +31% | 1 | 1 | 0% | 1,973 | 1,458 | -26% | 0 | 0 | — |
case-16 | fail→pass | 17,474 | 20,373 | +17% | 1 | 1 | 0% | 1,519 | 3,981 | +162% | 0 | 0 | — |
case-17 | fail→pass | 19,671 | 14,720 | -25% | 1 | 1 | 0% | 2,614 | 2,104 | -20% | 0 | 0 | — |
case-18 | pass→pass | 4,002 | 9,227 | +131% | 1 | 1 | 0% | 854 | 2,055 | +141% | 0 | 0 | — |
case-19 | fail→pass | 5,509 | 6,513 | +18% | 1 | 1 | 0% | 1,112 | 1,875 | +69% | 0 | 0 | — |
case-20 | fail→pass | 15,595 | 6,080 | -61% | 1 | 1 | 0% | 1,901 | 1,839 | -3% | 0 | 0 | — |
case-21 | fail→fail | 11,252 | 6,467 | -43% | 1 | 1 | 0% | 597 | 1,295 | +117% | 0 | 0 | — |
case-22 | fail→pass | 17,397 | 10,099 | -42% | 1 | 1 | 0% | 3,319 | 2,682 | -19% | 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 17 counted toward the lift figure. The other 5 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 +73 percentage points is the difference between those two pass rates over the 17 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.