Loading ToolBench
A mathematical calculator MCP server with tools for basic arithmetic, trigonometry, advanced math functions, and document automation (Word integration, email reporting)
This server has severe definition quality issues. 22 tools with mostly trivial descriptions (10-30 chars), no parameter descriptions beyond type hints, and inconsistent schemas. Math tools (add, subtract, etc.) have minimal docstrings. Three tools (open_word, draw_word_rectangle_with_text, send_email_report) lack proper input schemas or descriptions. The 'mine' tool has no explanation of what 'special mining' means. Image/string tools have vague purposes. No error handling guidance, no output schema documentation, and no indication of which tools are safe to retry. Naming is mostly acceptable (verb_noun pattern), but descriptions fail the 10-1024 character baseline and lack context for LLM selection.
Add two numbers
Add all numbers in a list
Cube root of a number
cos of a number
Create a thumbnail from an image
Divide two numbers
Draw rectangle and add text inside that rectangle in Word document
factorial of a number
Descriptions critically short (10-35 chars), below 194-char baseline. LLMs cannot determine when to select tools. E.g., 'sin of a number', 'log of a number', 'tan of a number' lack context for selection vs alternatives.
'mine' tool has vague, unexplained purpose ('special mining tool'). LLM cannot infer what it does or when to use it. Rename to reflect actual operation (e.g., 'subtract_twice') and document the formula.
No output schemas documented. LLMs do not know what fields to expect from responses, preventing downstream tool chaining and data extraction. E.g., create_thumbnail returns Image but no schema for structure.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 46 | <=2025-11-25 | v2 |
Return the first n Fibonacci Numbers
Return sum of exponentials of numbers in a list
log of a number
special mining tool
Multiply two numbers
Open Microsoft Word and create new document
Power of two numbers
remainder of two numbers divison
Sends an email report with execution logs and final result
sin of a number
Square root of a number
Return the ASCII values of the characters in a word
Subtract two numbers
tan of a number
No error handling guidance. Tools like divide (division by zero), log (negative input), factorial (large input) will fail silently. No recovery hints for LLM.
Word/email tools (open_word, draw_word_rectangle_with_text, send_email_report) lack clear descriptions of side effects. No indication these are WRITE operations with irreversible consequences. Agents cannot reason about retry safety.