{"id":66803,"date":"2026-06-09T00:25:18","date_gmt":"2026-06-09T00:25:18","guid":{"rendered":"https:\/\/www.europesays.com\/ai\/66803\/"},"modified":"2026-06-09T00:25:18","modified_gmt":"2026-06-09T00:25:18","slug":"metamask-arms-ai-agents-with-a-wallet-built-to-survive-mistakes","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ai\/66803\/","title":{"rendered":"MetaMask Arms AI Agents With A Wallet Built To Survive Mistakes"},"content":{"rendered":"

MetaMask has launched Agent Wallet, a self-custodial<\/a> product that lets AI agents<\/a> trade across decentralized finance while their owners keep control of spending limits and approvals.<\/p>\n

Key Points:<\/p>\n

MetaMask released Agent Wallet, a self-custodial tool built for autonomous AI agents.
\nThe wallet runs transactions through security checks, including simulation and threat scanning.
\nRoughly 200 users joined an early access program, with wider availability set for summer.<\/p>\n

MetaMask Opens Agent Wallet To Early Testers<\/p>\n

The company unveiled<\/a> the wallet on Monday, pitching it as a way for AI agents to autonomously swap tokens, trade perpetuals, and tap decentralized finance protocols across Ethereum-compatible chains and Hyperliquid. Roughly 200 users have joined a command-line Early Access Program so far, and the company expects a wider public rollout later this summer.<\/p>\n

Many projects on the market today instead hand agents direct access to private keys, a setup that can trigger unintended transactions or quietly drain user funds through plain error rather than outright hacks.<\/p>\n

Zhen Yu Tong, MetaMask’s senior director of product, said the infrastructure question cannot wait while agents already move and spend real money. “It’s genuinely day one for agents, but the infrastructure decision can’t wait,” he told reporters. Tong also cautioned that prompt injection, where rogue instructions hijack a model, remains an unsolved research problem rather than a bug that gets patched once.<\/p>\n

Also Read: Syscoin Halts Bridge After Exploit Spawns 5B Unauthorized SYS<\/a><\/p>\n

Guard Mode And Beast Mode Set The Limits<\/p>\n

In the default Guard Mode, owners set spending caps and approve which protocols an agent may touch, and any transaction that breaks those rules or looks suspicious requires two-factor authentication, backed by up to $10,000 in built-in protection. A looser Beast Mode lets agents act with more independence, yet a flagged malicious transaction still forces that second check no matter the operating mode chosen.<\/p>\n

Behind both modes, the wallet runs activity through transaction simulation, scam and malicious-contract detection, Blockaid-powered scanning, and Servo MEV protection, while private keys stay inside a hardware-isolated enclave during signing. Consensys founder Joe Lubin framed<\/a> the launch as a step toward machines that transact and verify one another on crypto rails, where protocols suit autonomous actors. “The next great expansion of the onchain economy won’t be driven by humans alone,” he said.<\/p>\n

Crypto Firms Race To Build Agent Wallets<\/p>\n

MetaMask’s move follows Coinbase, which introduced its own agentic wallet in February, and MoonPay, which added hardware-wallet support for human-approved agent transactions in March. The wallet also plugs into agent frameworks such as OpenAI’s Codex, Anthropic’s Claude Code, and Cursor, using Cubist’s enclave technology to keep key material beyond MetaMask’s own reach.<\/p>\n

Rival wallets and frameworks have spent recent months racing to keep private keys out of an agent’s reach, a contest that sharpened as developers handed bots growing control over portfolios, live trades, and direct ties to decentralized applications.<\/p>\n

Read Next: Bitcoin Pushes Above $63,000, Trimming Losses From A Savage Selloff<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"MetaMask has launched Agent Wallet, a self-custodial product that lets AI agents trade across decentralized finance while their…\n","protected":false},"author":2,"featured_media":66804,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[24,405,7537,37600],"class_list":["post-66803","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-ai","tag-ai-agents","tag-artificial-intelligence-agents","tag-metamask"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/66803","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=66803"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/66803\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/66804"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=66803"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=66803"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=66803"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}