excel_describe_sheets returns empty usedRange for files without <dimension> tag

[ttnyt8701]

Description

excel_describe_sheets returns an empty usedRange for Excel files that lack the <dimension> XML tag in their worksheet XML (xl/worksheets/sheet*.xml), even though the file contains valid data.

Steps to Reproduce

1. Obtain an Excel file exported from certain tools/systems that do not include the <dimension> tag (e.g., some Japanese business software exports)
2. Call excel_describe_sheets with the file

Expected Behavior

usedRange should reflect the actual data range (e.g., "A1:Z386").

Actual Behavior

{
  "backend": "excelize",
  "sheets": [
    {
      "name": "シート1",
      "usedRange": "",
      "tables": [],
      "pivotTables": [],
      "pagingRanges": []
    }
  ]
}

usedRange is empty, and subsequent excel_read_sheet calls also fail with "Invalid argument: no range available to read".

Root Cause

The <dimension> element is optional per the OOXML spec (ISO/IEC 29500). However, the excelize backend appears to rely on it for usedRange calculation. When the tag is absent, usedRange is returned as empty.

File with the issue (no <dimension> tag):

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<worksheet xmlns="http://schemas.openxmlformats.org/spreadsheetml/2006/main" ...>
  <!-- No <dimension> element here -->
  <sheetViews>...</sheetViews>
  <sheetData>
    <row r="1">...</row>
    <!-- 386 rows of data -->
  </sheetData>
</worksheet>

After re-saving with openpyxl (<dimension> tag added):

<dimension ref="A1:Z386"/>

→ usedRange correctly returns "A1:Z386".

Workaround

Re-saving the file with openpyxl (Python) before passing it to the MCP server adds the missing <dimension> tag:

import openpyxl
wb = openpyxl.load_workbook("file.xlsx")
wb.save("file.xlsx")

Environment

• excel-mcp-server: v0.12.0 (also confirmed on latest)
• Backend: excelize
• OS: Linux (Docker container)