Loading ToolBench
A LangChain-based conversational AI application with tool integration for mathematical operations and weather information
This server has severe definition quality issues. All 4 tools lack parameter descriptions (critical violation). Tool descriptions are extremely short (10-30 chars, well below the 34-char p10 baseline). The 'add' tool has a logic bug (multiplies instead of adds). No output schemas are documented. Parameter types are present but descriptions are null. This is a learning/demo project, not production-ready.
add two numbers
Answer to life, universe and everything!
subtract two numbers
Check how is the weahter in Fortaleza
All parameters lack descriptions. 'a' and 'b' in add/subtract have null descriptions, LLMs cannot infer what these represent or their valid ranges.
Tool descriptions are extremely short (10-30 chars). 'weather' and 'answer' have no input schema at all (empty objects).
No output schemas documented. LLMs cannot plan downstream calls or extract structured data from responses. All tools return untyped strings.
Logic bug in 'add' tool: implementation multiplies (a * b) instead of adding (a + b). This will cause incorrect results and agent confusion.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 22 | 2026-07-28+ | v2 |
Tool names lack action verbs. 'weather' and 'answer' are nouns, not verbs. Should be 'get_weather_fortaleza' and 'get_answer_to_everything' to signal intent.