A Model Context Protocol server for Excel file manipulation with capabilities including reading, writing, worksheet management, cell operations, formula handling, data formatting, and chart creation
This Excel MCP server has significant definition quality gaps across naming, descriptions, parameter documentation, and schema completeness. While 8 tools are declared with verb-based names (read_, write_, get_, add_, update_, apply_, format_, create_), the implementation exhibits critical deficiencies: (1) Parameter descriptions lack format/constraint details required by the rubric baseline of 72 chars average; (2) Output schemas are completely undocumented, no tool declares what it returns or which fields are available for chaining; (3) Generic, minimal descriptions fail the LLM optimization guideline of 50-200 chars with specific context; (4) No error handling guidance in any tool description; (5) The 'format_options' parameter in format_cells accepts an untyped object with no enumeration of allowed properties (bold, font_size, background_color, etc.). The source code shows tools are registered with @mcp.tool() decorators and basic parameter type hints, but schema details visible in the code are sparse. For example, read_excel_file accepts 'range' but does not specify A1:B5 notation or validation rules. The examples resource hints at usage but does not appear in tool descriptions themselves. Based on the rubric's baseline that 100% of A+ tools have documented return types and 94% have tool docstrings, this server falls well below production quality.
Add a new worksheet to an existing Excel file
Apply a formula to a cell in an Excel file
Create a chart in an Excel file
Apply formatting to cells in an Excel file
Get information about worksheets in an Excel file
Read data from an Excel file
Update a specific cell in an Excel file
No output schemas documented. None of the 8 tools declares what fields, types, or structure they return. LLMs cannot plan downstream tool calls or extract necessary chaining IDs (e.g., sheet_name after add_worksheet). This violates the rubric baseline that 100% of A+ tools have documented return types.
format_cells 'format_options' parameter is an untyped object with no enumeration or documentation of allowed keys (bold, font_size, background_color, etc.). This forces LLMs to guess valid property names and violates the constrained-input pattern.
Parameter descriptions lack format/constraint details. E.g., 'range' in read_excel_file says '(optional)' but does not specify A1:B5 notation, whether it supports named ranges, or what happens if invalid. Rubric baseline: parameter descriptions average 72 chars with format details. These average <30 chars.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 39 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 0 | - | v1 |
Write data to an Excel file
Tool descriptions do not include recovery guidance or dependency hints. E.g., get_worksheet_info should say 'Call this first to discover available sheets before using read_excel_file with sheet_name'. None mention what data is required or when to call them relative to other tools.
No error handling guidance. Tools raise generic exceptions without actionable recovery messages. E.g., if 'Sheet1' does not exist in read_excel_file, the LLM receives an error but no suggestion to call get_worksheet_info first or example sheet names to try.
create_chart 'chart_type' parameter does not constrain values to the documented list (line, bar, pie, scatter). Should be an enum to prevent LLMs from inventing invalid chart types.
Tool descriptions are too generic and do not explain WHEN to use each tool or how it differs from similar tools. E.g., 'Apply a formula to a cell' gives no context about formula syntax, cross-sheet references, or error conditions (circular deps, syntax errors).