Loading ToolBench
A Model Context Protocol server that provides domain analysis tools including WHOIS lookups, DNS record queries, and DNS issue detection and validation
Server provides 4 tools with complete JSON schemas and reasonable descriptions. All tools are READ_ONLY domain analysis operations. Naming follows verb_noun pattern (whois_lookup, dns_records, dns_health_check, domain_analysis). However, descriptions are generic and lack actionable context for LLM tool selection. No output schemas are documented, no error recovery guidance is provided, and parameter constraints are minimal. The server lacks pagination, result limits, and field descriptions that would enable proper tool composition. Baseline of 194 chars for tool descriptions is not met, most descriptions are 50-100 chars and lack specificity about when to use each tool vs alternatives.
Analyze DNS configuration for common issues
Get DNS records for a domain
Comprehensive domain analysis including WHOIS, DNS, and health checks
Get WHOIS information for a domain
No output schemas documented for any tool. LLMs cannot know what fields to expect in responses, forcing them to parse unstructured JSON blindly and preventing tool chaining.
Tool descriptions are too generic and lack context for selection. 'Get WHOIS information for a domain' (32 chars) tells the LLM WHAT but not WHEN to use this tool instead of dns_records or domain_analysis. Baseline tool descriptions average 194 chars; these average 58 chars and omit usage guidance.
No error handling guidance. Code catches exceptions but returns raw error strings ('Error performing WHOIS lookup for example.com: [Network error]'). LLM cannot distinguish retryable from permanent failures, cannot self-correct, and has no next-step guidance.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-21 | D | 51 | 2026-07-28+ | v2 |
dns_records accepts a 'record_types' array parameter but no parameter description explains valid values, format, or constraints. Description says 'DNS record types to query (A, AAAA, MX, NS, TXT, CNAME, SOA)' but this is in the parameter hint, not a proper description, and LLMs may hallucinate other record types.
Tool composition broken: domain_analysis is redundant with whois_lookup + dns_records + dns_health_check. No guidance on when to call which. LLM may waste tokens calling both domain_analysis AND its components, or may skip the combined tool not knowing it exists.
No pagination or result limits documented. dns_records may return hundreds of records; whois_lookup may return verbose structured data. Code does not cap results. Large responses waste tokens and degrade LLM reasoning.