Excel orchestrate

Skill YuYY2004/excel-skills/claude/skills/excel-orchestrate

18 Excel processing skills for Claude Code & Codex. XML direct ops for large files — 4-10x faster. Available in Chinese and English.

Install

npx -y skills add YuYY2004/excel-skills --skill excel-orchestrate

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

25 days oldThe repository was created 25 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

**Default entry point — all Excel file operations should consider this skill first.** ## Auto-trigger Conditions (any one suffices) / 自动触发条件（满足任一即触发） ### Condition 1: User mentioned an Excel file / 条件一：用户提到了 Excel 文件 As long as an Excel file (.xlsx / .xlsm / .csv / .xls / spreadsheet / workbook / worksheet) appears in context, AND the user is describing an operation to perform on it (modify/format/delete/merge/split/sort/dedup/filter/date/insert etc.), trigger this skill first — even if only one operation is described. ### Condition 2: Multi-step operations / 条件二：多步骤操作 - Multiple operations in one sentence - Delimiter-separated requirements - Parallel phrasing: "first...then...finally..." - Connective words: "also" "additionally" "and" "meanwhile" "at the same time" ### Condition 3: Vague/summary commands / 条件三：模糊/总结性指令 - "Process this file" "Handle this" "Take care of this" "Organize this" ## Non-trigger exceptions / 不触发的例外 - Pure query/scout: just asking "how many rows" "are there nulls" → directly use excel-scout - User explicitly names a skill → directly use that skill Trigger keywords (not limited to): ".xlsx" ".csv" "spreadsheet" "Excel" "process" "batch" "full workflow" "all" "organize" 触发词（不限于此）：".xlsx"".csv""表格""Excel""处理一下""批量""整个流程""全部""搞定""整理""先…再…""然后""还有""另外""并且""以及""同时""要求如下""以下几项"。 Workflow: Confirm it's an Excel operation → Call excel-scout for reconnaissance → Output report → User confirms → Execute one by one. 工作流程：确认是Excel操作 → 调 excel-scout 勘察 → 出报告 → 用户确认 → 逐个执行。

SKILL.md

5.9 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it

This skill is a meta-skill — it does not directly operate on Excel, but orchestrates the existing 16 skills. Follows the "Decompose→Report→Confirm→Execute" four-step approach. 本技能是元技能——不直接操作 Excel，而是编排已有的 16 个技能。遵循"拆分→报告→确认→执行"四步。

Excel Task Orchestration / Excel 任务编排

Workflow / 工作流程

User Complex Requirements / 用户复杂需求
    │
    ▼
[0] Scout First / 勘察先行 → Call excel-scout to locate target columns, show current values
    │
    ▼
[1] Parse Requirements / 解析需求 → Match scout results, decompose into independent sub-tasks
    │
    ▼
[2] Match Skills / 匹配技能 → Each sub-task maps to an existing skill
    │
    ▼
[3] Output Report / 输出报告 → Table showing execution plan (order, skill, input, expected output)
    │
    ▼
[4] User Confirms / 用户确认 → After confirmation, call skills one by one
    │
    ▼
[5] Summarize Results / 汇总结果 → Report status after each step completes

已有技能速查 / Available Skills Quick Reference

分类	技能	能干什么
🔍 查	`scout`	操作前勘察——定位目标列、展示当前值
	`find-duplicates`	扫描重复行
	`filter`	按条件筛选行
	`validate`	数据质量检查
🗑️ 删	`delete`	删除行/列
	`deduplicate`	去重
✏️ 改	`insert`	插入行/列
	`replace`	查找替换
	`mapping-replace`	按映射表批量替换
	`regex-clean`	正则清理（提取/删除/替换）
	`date-to-text`	日期转文本
	`format`	设置格式
📊 整	`sort`	排序
	`split`	按列拆分文件
	`merge`	合并多个文件
	`freeze`	冻结窗格
📋 基	`safe-workflow`	安全四步法

拆分原则 / Decomposition Principles

依赖关系：合并/筛选在前，改值/格式化在后
文件来源：先搞定输入（merge），再处理内容
同类合并：多个 mapping-replace 可以一次 XML 处理完成
大文件优先 XML：>10MB 的行操作自动走 XML 方案
不可逆操作靠后：删除类操作确认后再执行

报告格式 / Report Format

任务拆分报告 / Task Decomposition Report
═══════════════

原始需求: [用户原话]

执行计划:
| 序号 | 步骤 | 技能 | 输入 | 预期输出 |
|------|------|------|------|------|
| 1 | 合并28个文件 | excel-merge | 1000万-5000万/ | 单一xlsx |
| 2 | 日期列插文本 | excel-date-to-text | 合并结果 | 6个新列 |
| 3 | 类型→数字 | excel-mapping-replace | 合并结果 | 列值替换 |
| 4 | 国别→代码 | excel-mapping-replace | 合并结果 | 列值替换 |
| 5 | 新兴产业清理 | excel-regex-clean | 合并结果 | 列值清理 |

预计耗时: ~30分钟
风险点: 合并时需处理sharedStrings引用断裂

确认执行? (yes/no)

执行 / Execution

用户确认后，按顺序执行每步。每步完成后报告状态：

[1/5] excel-merge... done (451s, 333669行)
[2/5] excel-date-to-text... done (771s, 6列)
[3/5] excel-mapping-replace... done (697s)
[4/5] excel-mapping-replace... (已合并到第3步)
[5/5] excel-regex-clean... (已合并到第3步)

全部完成 / All Complete, 输出 / Output: 测试文件/终极合并.xlsx

注意事项

先报告再执行：拆分结果必须让用户看到并确认
步骤间传递文件：上一步的输出是下一步的输入，自动管理文件路径
同类操作合并：多个 mapping-replace 和 regex-clean 可一次 XML 处理，减少读盘次数
失败处理：某步失败时暂停，删除损坏产物，从备份恢复，询问用户是否继续或修改方案
sharedStrings 警告：如果涉及 merge + mapping-replace，提醒 sharedStrings 转换问题
每步自动备份：每个子任务执行前自动备份（时间戳命名），全部成功后仅保留最新3份