Loading ToolBench
TypeScript SDK and Claude Code / MCP plugin for r0x, the official native x402 facilitator on Robinhood Chain.
r0x-os exposes 18 blockchain/DeFi tools with consistent naming (verb_noun pattern) and reasonable descriptions (avg ~100 chars). All tools have input schemas with typed parameters. However, descriptions lack LLM-optimization guidance (e.g., when to use vs similar tools, prerequisites, recovery hints). Parameter descriptions are minimal (5-20 chars) and lack format/constraint details. Output schemas are not documented. Error handling is absent from descriptions. No tool annotations (readOnlyHint/destructiveHint). Risk classification exists but not surfaced in descriptions.
ETH + USDG balance for any Robinhood Chain address. Costs $0.01 USDG.
Across Protocol bridge quote for moving USDC or WETH from Ethereum, Arbitrum or Base onto Robinhood Chain. Costs $0.01 USDG.
Signs an unsigned transaction (from the send tool) with a private key and broadcasts it on Robinhood Chain. Costs $0.01 USDG. Defaults to the wallet configured via r0x_setup if no privateKey is given. Use with care, this moves real funds.
Sends a message to the r0x agent and returns its reply. Costs $0.01 USDG.
Balance plus step-by-step funding instructions for a Robinhood Chain address. Costs $0.01 USDG.
Mints a Uniswap V3 liquidity position for any token pair, with a custom price range. Costs $0.01 USDG. If this worker was configured with a signer, the position is minted automatically and a real tx hash is returned.
Parameter descriptions are minimal (5-20 chars) and lack format/constraint guidance. E.g., 'address' param in balance/fund/liquidity_positions lacks format hint (0x..., length). 'token' in price/trade lacks enum constraint or valid values list. 'slippageBps' in trade lacks range (e.g., 1-10000). LLMs cannot infer constraints from names alone.
No output schemas documented. Tools return complex objects (tx details, swap quotes, liquidity positions) but LLMs have no visibility into response structure. This forces agents to guess field names and risks extraction errors. E.g., does 'broadcast' return {txHash, status, gasUsed} or {hash, receipt, ...}?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
Lists every Uniswap V3 liquidity position NFT owned by an address. Costs $0.01 USDG.
Withdraws and collects a Uniswap V3 position by tokenId. Set burn to also destroy the now-empty position NFT. Costs $0.01 USDG. If this worker was configured with a signer, the position is removed automatically and a real tx hash is returned.
Uniswap V3 pool reserves and implied price for any token pair on Robinhood Chain. Costs $0.01 USDG.
Current USD price for ETH or USDG on Robinhood Chain. Costs $0.01 USDG.
Live Uniswap V3 swap price quote for any token pair on Robinhood Chain, no transaction constructed. Costs $0.01 USDG.
Initialize or update the r0x wallet and optional spend limit
Constructs an ETH or USDG transfer on Robinhood Chain. Costs $0.01 USDG. If this worker was configured with a signer, the transfer is broadcast automatically and a real tx hash is returned.
View or update the USD spend limit for paid skill calls
Swaps any two tokens on Robinhood Chain via Uniswap V3, pass "ETH", "USDG", or any ERC20 address, there is no allowlist. Costs $0.01 USDG. If this worker was configured with a signer, the swap (and any approval it needs) is broadcast automatically and real tx hashes are returned.
Decoded transaction details for any Robinhood Chain transaction hash. Costs $0.01 USDG.
Generates a fresh Robinhood Chain wallet keypair server-side. Costs $0.01 USDG.
Live Morpho steakUSDG vault APY, the vault behind Robinhood Earn. Costs $0.01 USDG.
Descriptions lack LLM-optimization guidance. No 'when to use' context, no dependency hints, no recovery guidance. E.g., 'chat' tool description does not explain when to call it vs direct agent reasoning. 'wallet' does not clarify if it's for generating new wallets or querying existing ones. Descriptions average ~100 chars but lack actionable structure.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Risk classification exists in metadata (READ_ONLY, WRITE, IRREVERSIBLE) but is not exposed via MCP tool annotations. LLMs cannot distinguish safe tools from destructive ones without explicit hints. 'broadcast' and 'trade' are IRREVERSIBLE but lack destructiveHint.
Error handling not documented in tool descriptions. No guidance on retryability, user-fixable errors, or recovery steps. E.g., 'broadcast' costs $0.01 USDG but description does not explain what happens if wallet lacks funds or if tx fails on-chain. Agents have no recovery path.