Loading ToolBench
A multi-protocol agent system that combines A2A (Agent-to-Agent) protocol with MCP (Model Context Protocol) tools for numerical data analysis. Provides a Chainlit UI orchestrator that routes user queries to specialized agents, with a Data Analysis agent powered by LangGraph and local Ollama LLM.
Two tools with minimal definitions. Both have short descriptions (under 20 chars baseline threshold) and basic parameter schemas. No output schemas documented. No error handling guidance. No tool annotations. Naming is verb-based but generic. Parameter descriptions exist but are minimal. No composition guidance or chaining support. This server falls well below production baseline (median ~50).
Add two numbers.
Multiply two numbers.
Tool descriptions are too short (16 chars each). Baseline is 34-392 chars; <20 chars cannot guide LLM selection. 'Add two numbers.' and 'Multiply two numbers.' lack context on when to use, what they return, or prerequisites.
No output schemas documented. LLMs cannot plan downstream calls or extract results. Both tools return float but this is not declared in the tool definition.
No error handling guidance. If a call fails (e.g., invalid input type), LLM has no recovery path. No actionable error messages defined.
No tool annotations (readOnlyHint, idempotentHint). Both tools are read-only and idempotent but this is not declared in the schema.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 41 | 2026-07-28+ | v2 |
Parameter descriptions are minimal ('First number', 'Second number'). Baseline is 72 chars avg. No format constraints, ranges, or type hints beyond the JSON Schema type field.