Loading ToolBench
Medical terminology intelligence for AI agents — ICD-10 lookup, MedDRA hierarchy, RxNorm drug concepts, SNOMED CT, and cross-terminology mapping.
Server has 8 well-intentioned medical terminology tools with clear domain focus. All tools have descriptions (194 - 250 chars, within baseline 34 - 392 range) and input schemas with types. However, critical gaps reduce quality: (1) Output schemas are NOT documented, responses are free-text markdown blocks, not structured objects. LLMs cannot parse results reliably or chain tools. (2) No error handling guidance, tools fail silently or return generic 'not found' messages without recovery hints. (3) Parameters lack validation constraints, 'query' and 'term' accept any string; no regex, length limits, or format guidance. (4) Tool names are inconsistent: 'lookup_*' vs 'map_*' creates cognitive load; 'cross_terminology_map' is vague (does it search or transform?). (5) Incomplete source code, index.js truncated mid-function, cannot verify full implementation. Average per-tool score: 62/100.
Map a medical concept across multiple terminologies (ICD-10 ↔ MedDRA ↔ SNOMED CT ↔ RxNorm). Returns all equivalent codes and terms. Essential for data integration, EHR interoperability, and regulatory submissions.
Look up Common Terminology Criteria for Adverse Events (CTCAE) grade definitions. Returns severity grades (1–5), clinical descriptions, and oncology context. Essential for cancer trial adverse event grading and reporting.
Search ICD-10-CM codes by diagnosis name or look up a specific code. Returns code, full description, and category. Essential for clinical documentation agents.
Look up a MedDRA Preferred Term (PT), identify its System Organ Class (SOC), hierarchy level, and related terms. Critical for adverse event coding, pharmacovigilance, and regulatory submissions.
Search RxNorm drug concepts by name or RxCUI. Returns drug name, strength, form, route, and NDF-RT classification. Essential for medication reconciliation, drug interaction checking, and clinical decision support.
Search SNOMED CT concepts by term or SNOMED ID. Returns concept ID, fully specified name (FSN), preferred term, and hierarchy. Essential for EHR integration, clinical data exchange, and semantic interoperability.
Output schemas not documented. Tools return free-text markdown blocks instead of structured JSON objects. LLMs cannot reliably extract fields or chain results to downstream tools.
No error handling guidance. Tools fail with generic 'not found' messages or API errors without recovery hints. LLMs cannot determine if error is retryable or requires user intervention.
Parameter validation constraints missing. 'query', 'term', 'concept' parameters accept any string with no length limits, regex patterns, or format guidance. Invites invalid input from LLMs.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2026-07-28+ | v2 |
Map an ICD-10 diagnosis code to likely MedDRA adverse event terms. Useful for converting clinical diagnoses to pharmacovigilance terminology for FAERS reporting and safety signal detection.
Map an RxNorm drug concept to ATC (Anatomical Therapeutic Chemical) classification. Returns ATC code, level hierarchy, and therapeutic category. Essential for drug classification, pharmacoepidemiology, and international drug safety reporting.
Inconsistent tool naming. 'lookup_*' (4 tools) vs 'map_*' (2 tools) vs 'cross_terminology_map' (1 tool) creates cognitive load. 'cross_terminology_map' name is vague, does not clearly signal it searches/transforms across terminologies.
Source code truncated. index.js cuts off mid-function ('const normalized = term.toLowerCase().trim();'). Cannot verify full implementation, error handling, or API integration completeness.