Model Context Protocol (MCP) implementation for AllThatNode blockchain RPC services
All 8 tools share identical, minimal schemas and descriptions. Tool names lack action verbs (ethereum_mainnet_rpc vs get_ethereum_mainnet_balance). Descriptions are generic ('JSON-RPC API for X'), under 30 chars, providing no context for when/why to use each tool. Parameters lack descriptions: 'method' and 'params' are documented at schema level but lack guidance on valid values, constraints, or examples. No output schema documented. No error handling guidance. Tools are pass-through RPC proxies with no agent-friendly abstraction, LLMs cannot infer what methods are available or what each returns without external knowledge.
JSON-RPC API for arbitrum-one
JSON-RPC API for arbitrum-sepolia
JSON-RPC API for ethereum-holesky
JSON-RPC API for ethereum-mainnet
JSON-RPC API for ethereum-sepolia
JSON-RPC API for optimism-mainnet
JSON-RPC API for optimism-sepolia
JSON-RPC API for sui-mainnet
Tool names lack action verbs. Names like 'ethereum_mainnet_rpc' do not convey intent (get, send, query, etc.). LLMs cannot infer what action occurs when calling the tool.
Descriptions are generic and under 30 characters ('JSON-RPC API for X'). No context on when to use each tool, what methods are available, or what data is returned.
Parameter 'method' and 'params' lack descriptions in the schema. LLMs cannot determine valid method names, parameter formats, or constraints without external documentation.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 24 | <=2025-11-25 | v2 |
No output schema documented. LLMs cannot plan downstream operations or extract required fields from RPC responses.
No error handling guidance. Tools return raw JSON-RPC errors without recovery hints. LLMs cannot self-correct or determine if errors are retryable.