Trust Wallet starts building the agent layer for self-custody

AI agents are beginning to trade, stake, and move funds autonomously. Have you ever wondered which wallet they’d use to sign transactions? Trust Wallet is making a strong case to be that wallet. One of the most widely used self-custody platforms in crypto is now pushing further into AI with the launch of the Trust Wallet Agent Kit, a developer toolkit that lets AI agents carry out wallet-connected crypto actions within permissions defined by the user. The release follows the launch of Trust Wallet’s developer portal last week, which gave AI agents read-only access to crypto data across more than 100 chains and set the stage for more interactive agent workflows. The move targets two groups at once: the developers building AI-powered crypto applications and the millions of users who may soon rely on AI agents to manage on-chain activity. AI tools for developers The first pieces of this strategy arrived at the end of February, when Trust Wallet released a set of AI-focused developer tools. One of them is Claude Code Skills, an open-source skills marketplace on GitHub that extends Anthropic’s Claude coding assistant with native knowledge of Trust Wallet’s architecture and APIs. Instead of developers constantly feeding documentation into AI assistants, the system already understands how Trust Wallet works. Five specialized skill modules are available at launch. These cover wallet creation and transaction signing across more than 100 blockchains, integration with Web3 providers such as Ethereum, Solana, Cosmos, and Bitcoin, token metadata management, Trust Wallet’s smart contract wallet (Barz), and developer utilities like deep links and WalletConnect integrations. Alongside this, Trust Wallet introduced an MCP server, which connects its entire developer documentation library directly to AI coding tools such as Claude, Cursor, and VS Code. The goal is to make building AI-powered applications around wallets easier and faster. That push is now expanding beyond documentat...