Put a wallet in front of any x402-paid MCP server, and hold a metered line on the ones that sell time. Stock MCP clients cannot construct x402 payments; this one does, and proxies your existing client through it.
Two tools with minimal documentation. 'rpc' has a basic description and input schema with three parameters (chain, method, params), but lacks output schema documentation and error guidance. 'sample' has only a trivial description ('A free sample tool that demonstrates the bridge's catalog') with no input schema visible and no output structure defined. Neither tool provides recovery guidance, parameter constraints (enums for chain values), or examples of expected responses. Descriptions are too generic to guide LLM tool selection. No tool annotations (readOnlyHint, etc.) present. Tool definitions appear inferred from test files rather than explicit registration in main server code.
Execute JSON-RPC calls against the upstream MCP server's RPC endpoint
A free sample tool that demonstrates the bridge's catalog
No output schema documented for either tool. LLMs cannot plan downstream calls or extract required fields (e.g., does 'rpc' return {result, error, blockNumber}?). Violates pattern:tool requirement for documented return types.
'sample' tool has no visible input schema (empty object {}). Cannot verify parameter types or constraints.
'rpc' params array lacks enum constraint for 'chain' parameter. Should enumerate valid chains (base, mainnet, etc.) instead of free-form string. Invites hallucinated chain names.
No error handling guidance. If RPC call fails (invalid method, network error, rate limit), tool provides no recovery hint. LLM cannot self-correct or retry intelligently.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 46 | <=2025-11-25 | v2 |
'sample' description (18 chars) is below 20-char minimum and provides no actionable context. Does not explain when to call it, what it returns, or how it differs from 'rpc'.