Wrapped several on-chain data skills with uxc, then let the openclaw agent combine them for use. I tried it out simply, and the results were decent. Prompt: Help me find the most noteworthy on-chain trading opportunities in the last 24 hours on Base and Ethereum. First, do market discovery, then verify token activity and protocol growth, check for contract risks, and finally determine if participation is possible now, but do not execute any trades. Let me know the skills or tools used for the analysis. OpenClaw version: 2026.2.23 Model: MiniMax M2.1 See the results in the screenshot. uxc has provided 6 skills related to on-chain data: okx-mcp-skill Responsible for market discovery and trading feasibility analysis. Can check hot tokens, prices, transactions, candlesticks, holders, wallet assets, DEX liquidity, and quotes, suitable for making judgments like "what is currently being traded in the market" and "can this coin be bought now". etherscan-mcp-skill Responsible for contract, address, and transaction-level due diligence. Can check contract verification, source code, ABI, address holdings, token holders, transaction details, etc., suitable for checking "does this project/address have obvious risks". dune-mcp-skill Responsible for aggregation analysis and trend validation. Can find tables, write SQL, run queries, suitable for analyzing 7d/30d trading volume, active addresses, protocol usage trends, track comparisons, etc., answering "is this wave of popularity sustainable". thegraph-mcp-skill Responsible for protocol-level data validation. Queries users, swaps, fees, TVL, volume, and other entities and metrics through subgraphs, suitable for determining "is the rise of the token driven by real protocol growth". thegraph-token-mcp-skill ...