Loading ToolBench
In-memory vector store with TF-IDF vectorization and cosine similarity search.
Single tool with adequate schema and clear description, but lacks critical production patterns. Tool name 'data_vector_search' is verb-noun compliant. Description (194 chars) is within baseline range (p10=34, p90=392) and includes use-case guidance with negative examples. Input schema is complete with types and descriptions for all params. However, output schema is not formally documented in the tool definition, only visible in runtime response. No error handling guidance, no idempotent/destructive hints, no pagination support despite potential for large result sets. Missing tool annotations and recovery patterns.
Use this when you need to store text documents and search them by semantic similarity. Accepts documents to store and a query to search. Uses TF-IDF vectorization with cosine similarity to find the most relevant matches. Returns top-k results with similarity scores. Do NOT use for web search — use web_search_query instead. Do NOT use for keyword research — use keyword_research instead. Do NOT use for text classification — use text_classify instead.
Output schema not formally documented in tool definition. Only visible in runtime response (results array with text, similarity, index fields). LLMs cannot plan downstream operations without knowing response structure.
No error handling guidance. Tool returns 400 errors for missing query or empty namespace, but responses lack actionable recovery hints (e.g., 'Provide documents array to index first' is present but not structured as a recovery guide). LLMs cannot self-correct without explicit next-step guidance.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Tool is read-only but this is not declared in schema. Agents cannot determine safety classification without explicit hints.
topK parameter lacks explicit bounds validation in description. Schema shows 'number' type but description says 'default: 3, max: 10' without stating minimum. Code enforces Math.min/Math.max but LLMs cannot read implementation, they rely on description text.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 65 | 2025-06-18+ | v2 |
No pagination support despite in-memory store potentially holding hundreds of documents. Tool returns all top-K results but no cursor, offset, or limit mechanism for large result sets. Violates pattern:paginated-result baseline.