Model Context Protocol server for reading, writing, and manipulating XLSX/XLSM Excel files without Microsoft Excel. Provides comprehensive Excel automation capabilities for AI/LLM integration.
Critical gaps across all three tools. Tool names lack clear action verbs (enhanced_autofit_columns, apply_consistent_number_format, ensure_chart_spacing are vague about what they modify). Descriptions are present but generic (194-char baseline; these average ~150 chars but lack WHEN/WHY context). Most critically: input schemas expose openpyxl worksheet objects as parameters, this is a fundamental design flaw. LLMs cannot instantiate or manipulate Python objects; they need file paths, sheet names, or cell ranges. No output schemas documented. No error handling guidance. Parameters lack type validation hints (min_width/max_width are numbers but no range constraints stated). No idempotency guarantees. Tools violate the chat-data-model pattern by requiring internal API objects instead of human-friendly identifiers.
Apply consistent number formatting to all cells with numeric values. Auto-detects and formats currency and percentage values based on cell content and neighboring cells.
Enhanced auto-fit for all columns in a worksheet. Calculates optimal column width based on cell content, handles multi-line text with wrapping, and respects minimum and maximum width constraints.
Ensure proper spacing around charts to avoid overlaps. Checks for content in the chart area and adjusts position if needed to maintain visual clarity.
Input schemas require openpyxl worksheet objects as parameters. LLMs cannot instantiate or manipulate Python objects. Tools must accept file_path + sheet_name or cell_range instead.
No output schemas documented. LLMs cannot plan downstream calls or extract results. Each tool must declare what it returns (success/failure, affected cells, new dimensions, etc.).
Tool names lack action verbs or are ambiguous. 'enhanced_autofit_columns' does not clearly state it modifies column widths. 'ensure_chart_spacing' is passive. Use 'adjust_column_widths', 'format_numbers', 'reposition_chart'.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 35 | <=2025-11-25 | v2 |
Numeric parameters (min_width, max_width, chart_width, chart_height) lack range constraints in descriptions. LLMs may pass absurd values (width=999999). Add explicit min/max bounds.
No error handling guidance. If autofit fails (e.g., worksheet is empty), what does the tool return? Can the agent retry? Must it ask the user? Descriptions must include recovery paths.