Loading ToolBench
Local Agent Discovery for Agent-to-Agent (A2A) protocol - provides mDNS/DNS-SD advertisement, discovery endpoints, and A2A AgentCard serving
This is a LAD-A2A reference implementation, not a traditional MCP server. Tools are HTTP endpoints (FastAPI routes) without explicit MCP tool registration, schemas, or descriptions visible in the provided code. Of 8 tools listed, only 2 have partial input schemas shown (hotel_a2a_jsonrpc, aria_websocket_chat); the rest lack any visible schema definition. Descriptions are present but generic (e.g., 'LAD-A2A discovery endpoint'). No parameter descriptions, no output schemas, no error handling guidance. Tool names lack action verbs (lad_discovery, agent_card, health_check are nouns/descriptors, not verb_noun patterns). This implementation appears to be a protocol reference rather than a production-grade MCP server.
A2A AgentCard endpoint - serves the full A2A AgentCard with agent metadata, capabilities, and authentication requirements
User Agent (Aria) discovery endpoint - discovers local A2A agents via mDNS and well-known endpoints
User Agent (Aria) WebSocket endpoint - handles real-time chat with agent discovery, connection, and A2A message routing
Health check endpoint - returns server status, agent name, and feature availability (mDNS, TLS, signing)
Hotel Agent A2A JSON-RPC 2.0 endpoint - handles SendMessage and other A2A methods for hotel concierge queries
Hotel Agent A2A AgentCard endpoint - serves hotel concierge agent metadata with skills for hotel info, dining, spa, local recommendations, and concierge services
Hotel Agent LAD-A2A discovery endpoint - returns hotel concierge agent information
No explicit MCP tool registration visible. Tools appear to be HTTP endpoints, not MCP-registered tools with schemas and metadata.
Tool names lack action verbs. Names like 'lad_discovery', 'agent_card', 'health_check' are nouns/descriptors, not verb_noun patterns (e.g., 'get_agents', 'fetch_agent_card'). LLMs cannot infer intent from these names.
6 of 8 tools have no visible input schema. Only hotel_a2a_jsonrpc and aria_websocket_chat show partial schemas. Without schemas, LLMs cannot validate inputs or understand parameter types.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 30 | <=2025-11-25 | v2 |
LAD-A2A discovery endpoint - returns list of discoverable agents on the network
No output schemas documented. Descriptions do not explain what fields are returned or how to chain results to downstream tools.
No error handling guidance. Descriptions do not explain failure modes, recovery steps, or what the LLM should do if a call fails.