Loading ToolBench
Convert Markdown to Word (.docx) with Chinese government document formatting (公文格式规范). Supports custom format specs, reference templates, media extraction, and optional auto-generation of modification suggestions. Also converts Markdown tables to Excel (.xlsx).
Three tools with complete schemas and detailed descriptions. Tool names follow verb_noun pattern (convert_*, analyze_*). Descriptions are comprehensive (200-400 chars) and explain formatting behavior well. However, parameter descriptions lack actionable constraints (no enums for format choices, no min/max for numeric params). Error handling returns user-friendly messages but lacks recovery guidance. Output schema not formally documented. No tool annotations (readOnlyHint/destructiveHint). All tools marked WRITE risk but no confirmation/dry-run pattern.
Analyze a Markdown document's structure and content, generating a structured modification suggestions report (xx-修改建议.md) covering heading hierarchy, required sections, format issues, content problems, list continuity, and Chinese punctuation.
Convert a Markdown file to Word (.docx) with formatting. Applies Chinese government document formatting (公文格式规范) by default: - Headings: 黑体 (H1 三号16pt, H2 小三15pt, H3 四号14pt) bold, left-aligned - Body: 仿宋_GB2312 四号14pt, justified, 2-char first-line indent - Tables: 仿宋_GB2312 五号10.5pt, center-aligned - Line spacing: fixed 28pt for body/headings, single for tables
Convert Markdown tables to an Excel (.xlsx) file. Extracts every HTML <table> and Markdown pipe table from the document (in document order) and writes each to its own worksheet. Preserves colspan/rowspan merged cells, inline <b> bold, and text-align alignment. Styling: - Font: 微软雅黑 (11pt body), thin borders on all cells - Header cells (HTML <th> or the first row of a pipe table): bold + light gray fill - All cells: vertical center, wrap text - Column widths and row heights estimated from content
No tool annotations (readOnlyHint/destructiveHint/idempotentHint). All three tools perform file I/O with side effects but lack destructiveHint=true to signal to agents that these are irreversible operations.
Output schema not documented. convert_markdown_to_docx returns a string summary; convert_markdown_to_xlsx and analyze_and_suggest return values are inferred but not formally specified. LLMs cannot plan downstream operations without knowing return structure.
Parameters lack actionable constraints. 'format_spec' and 'reference_docx' accept arbitrary paths with no validation hints. 'body_size' (integer) has no min/max bounds. No enums for mutually exclusive options (skip_format vs format_spec). LLMs cannot self-validate inputs.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | 2026-07-28+ | v2 |
Error handling returns user-friendly Chinese messages but lacks recovery guidance. 'PandocNotFoundError' returns install link (good), but other errors like 'ConversionError' provide no next steps. No categorization of retryable vs fatal errors.
No confirmation/dry-run pattern for destructive operations. convert_markdown_to_docx and convert_markdown_to_xlsx overwrite output files without warning. Agents cannot preview changes before committing.