Loading ToolBench
Intent-first search that finds the right tool from a vague description, for people and AI agents (UI, HTTP API and MCP).
Intently exposes 8 tools via HTTP with partial schema documentation. Most tools have descriptions (50-150 chars), but parameter documentation is inconsistent. Input schemas are visible for all tools, but lack depth: many parameters missing type constraints, enums, or validation guidance. No output schemas documented. Error handling is absent, tools return raw results with no recovery guidance. Tool naming follows verb_noun convention (search, find-tools, answer, entity, categories, category, health, mcp), which is good, but descriptions are generic and lack WHEN/WHY context. No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Composition is reasonable, tools are single-purpose, but chaining support is unclear (no documented output fields for downstream tool inputs).
Short cited answer over the top search results
Category folders with counts of tools in each category
Top tools in a specific category
Everything about one tool: links, mentions, related tools, versions
Tools for a task, grouped by capability. Decomposes a task into distinct capabilities and finds matching tools for each.
Index and LLM status, embedder readiness, available LLM keys
Model Context Protocol endpoint for stateless JSON responses over HTTP
No output schemas documented for any tool. LLMs cannot predict response structure, forcing them to parse unstructured results and plan downstream calls blindly.
Parameter descriptions lack actionable constraints. E.g., 'q' parameter has maxLength=20000 but no guidance on format, language, or expected query structure. 'llm' and 'rerank' parameters use 0/1 instead of boolean enum.
No error handling or recovery guidance. Tools return raw results with no indication of failure modes, retryability, or next steps if a search returns no results or an entity is not found.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
Intent search that finds tools matching a query with optional LLM-backed reranking
Tool 'mcp' has vague description ('Model Context Protocol endpoint for stateless JSON responses over HTTP') and unclear purpose. Parameter type is 'object' with no schema, LLMs cannot determine what fields to pass.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). All tools appear read-only (marked READ_ONLY in metadata), but this is not declared in tool definitions, forcing LLMs to infer safety.