Loading ToolBench
MCP server that provides access to Hokai tools-tree API and health check functionality
Both tools lack parameter descriptions entirely, have minimal descriptions (under 20 chars baseline), and expose no input schemas. get_tools_tree and scanner_status are READ_ONLY but descriptions do not state this. No error handling guidance, no output schema documentation, and no parameter validation. The server implements only 2 trivial tools with no composition or chaining capability. This is a minimal proof-of-concept, not production-grade tooling.
Fetch Hokai tools-tree JSON.
Simple health check for the MCP server.
No input schemas visible for either tool. Both tools accept empty input {} but no JSON Schema is declared or documented.
Tool descriptions are under 20 characters and lack context. 'Fetch Hokai tools-tree JSON' and 'Simple health check for the MCP server' do not explain WHEN to use each tool, what it returns, or how it differs from similar tools.
No output schema documented. Return types are dict but structure, field names, and nested objects are not specified. LLMs cannot plan downstream calls or extract data reliably.
No error handling or recovery guidance. get_tools_tree calls an external API with timeout=30 but does not document failure modes, retry logic, or what the LLM should do if the API is unreachable.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 36 | 2026-07-28+ | v2 |
Tool descriptions do not state that both tools are READ_ONLY. Agents need explicit indication of which tools modify state and which are safe to call repeatedly.