Excel 数据自动清洗引擎 Pro
ReviewAudited by ClawScan on May 10, 2026.
Overview
The Excel-cleaning workflow is mostly coherent, but it may use external web search on sensitive spreadsheet fields such as phone numbers without clear consent or data-sharing limits.
Use this only with files you intentionally provide, keep a backup of the original spreadsheet, and ask the agent not to use web_search or cloud export on personal, customer, financial, or ID data unless you explicitly approve the exact fields and destination.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Sensitive customer, employee, financial, or ecommerce data could be exposed to an external search service during cleaning.
The skill may send spreadsheet-derived phone-number data to an external search provider for validation, but it does not define approval, minimization, destination, or retention boundaries.
- `web_search` — 辅助数据验证（如手机号归属地）
Require explicit user confirmation before any web_search on spreadsheet contents, avoid sending full personal identifiers, and prefer local/offline validation rules where possible.
The agent may run Python commands on local files to clean and generate spreadsheets.
The skill tells the agent to run local Python processing. This is expected for Excel automation, but users should be aware that local code execution is part of the workflow.
- `exec` — 运行 Python openpyxl/pandas 处理
Run it only on intended files, keep backups of originals, and review generated outputs before relying on them.
If used, the skill could write cleaned data into a third-party business workspace.
The optional Feishu export implies delegated account or workspace write permission, although no credential contract is declared in the metadata.
"导出飞书表格" → 写入飞书多维表格
Use Feishu export only after an explicit request, confirm the target workspace/table, and use the least-privileged account or token available.