Crypto AI Platform Bankr Locks Down System After Hacker Breaches 14 Crypto Wallets

bitcoinistPublished on 2026-05-20Last updated on 2026-05-20

Abstract

Crypto AI trading platform Bankr has locked down its system after a hacker breached 14 user wallets. The attack, which targeted wallets automatically created via interactions with Bankr's AI bot on X, appears to be a social engineering scheme exploiting the platform's connection with Grok. Security experts suspect the use of "prompt injection" to manipulate the AI into approving unauthorized transactions. Losses from individual wallets reached as high as $150,000, with a total of $440,000 identified in three attacker-controlled addresses. Tech entrepreneur Austen Allred was among the victims. Bankr has halted all transactions, pledged full reimbursement for lost funds, and advised affected users to set up new wallets and transfer any remaining assets. This incident follows a series of major crypto exploits in recent months.

Tech entrepreneur Austen Allred was among the victims. His wallet, tied to a project called Kelly Claude AI assistant, was drained of Ether — though the hacker left his memecoin holdings untouched. Allred said there was no sign anyone else had logged into his Bankr account, suggesting the attacker got to the private keys through other means.

How The Attack Unfolded

Bankr is a crypto trading tool that lets users send plain-language instructions — like “swap this token” or “transfer funds” — to an AI that carries out the trades.

The platform also creates a crypto wallet automatically for every X account that interacts with its bot.

That feature had already drawn attention earlier this year, when someone reportedly tricked Grok into telling Bankr to launch a token, then pulled funds from it into a wallet they controlled.

Tuesday’s incident appears to follow a similar pattern. Yu Xian, founder of blockchain security firm SlowMist, said the breach was likely a social engineering scheme aimed at the AI agent.

ETHUSD now trading at $2,129. Chart: TradingView

According to Xian, the attacker exploited the trust connection between Grok and Bankrbot to push through unauthorized transaction approvals.

He identified three wallet addresses linked to the attacker that together held $440,000 in crypto.

Xian also pointed to prompt injection as part of the method — a technique where malicious instructions are fed to an AI to manipulate its behavior.

Bankr Pledges Full Reimbursement

Bankr confirmed the breach in a post on X, saying it had identified an attacker who accessed 14 wallets. The platform said it shut down all transaction activity — swaps, transfers, and token deployments — while the investigation continues. It also pledged to cover all losses.

Users were warned not to sign any transactions for now. For those with wallets already hit, Bankr told them to stop using the affected accounts entirely, set up a new wallet with a fresh seed phrase on a clean device, and transfer any remaining tokens or NFTs out immediately.

If assets can’t be moved, revoking existing approvals was advised. Bankr also flagged the possibility of malware, urging users to check their computers and phones for suspicious software or browser extensions.

What Users Lost

Some users reported losing as much as $150,000 from a single wallet. The exact total across all 14 breached wallets has not been confirmed.

The attack adds to a rough stretch for the crypto space. Bad actors stole more than $168 million in the first quarter of the year.

April brought two of the biggest hits so far — a $280 million exploit of Drift Protocol and a $292 million breach of Kelp.

Just a day before the Bankr incident, the Ethereum bridge of Verus Protocol was also reportedly hit.

Featured image from Unsplash, chart from TradingView

Related Questions

QWhat is Bankr and how does it function as described in the article?

ABankr is a crypto trading tool that allows users to send plain-language instructions, like 'swap this token' or 'transfer funds', to an AI which then carries out the trades. The platform also automatically creates a crypto wallet for every X (formerly Twitter) account that interacts with its bot.

QAccording to the article, what was the likely method used by the attacker to breach the Bankr wallets?

AAccording to Yu Xian, founder of SlowMist, the breach was likely a social engineering scheme aimed at the AI agent. The attacker exploited the trust connection between Grok and Bankrbot to push through unauthorized transaction approvals, and prompt injection—feeding malicious instructions to manipulate the AI's behavior—was part of the method.

QWhat actions did Bankr take in response to the security breach?

ABankr confirmed the breach, identified that an attacker accessed 14 wallets, and temporarily shut down all transaction activity including swaps, transfers, and token deployments while investigating. They pledged to reimburse all lost funds and warned users not to sign any transactions.

QWhat advice did Bankr give to users whose wallets were affected by the hack?

ABankr advised affected users to stop using the compromised accounts entirely, set up a new wallet with a fresh seed phrase on a clean device, and immediately transfer any remaining tokens or NFTs out. If assets couldn't be moved, they advised revoking existing approvals and checking devices for malware or suspicious software/extensions.

QBesides the Bankr incident, what other major crypto exploits were mentioned as part of a 'rough stretch' for the crypto space?

AThe article mentions that bad actors stole over $168 million in Q1 of the year. In April, there were two major exploits: a $280 million exploit of Drift Protocol and a $292 million breach of Kelp. Just a day before the Bankr incident, the Ethereum bridge of Verus Protocol was also reportedly hit.

Related Reads

Agent Race Ends, Super Workbench Takes Over

The era of fragmented AI agents is ending. Over the past month, China's tech giants—Tencent, Alibaba, and ByteDance—have simultaneously shifted strategy: instead of launching new, standalone AI agents, they are consolidating their various agent projects into unified "super workbenches." Tencent integrated its QClaw teams into WorkBuddy, a strategic product hailed as a potential third flagship after QQ and WeChat. Alibaba is merging its QoderWork, Wukong, and MuleRun agents into a new "Qianwen Office" platform under DingTalk's leadership. ByteDance rebranded its TRAE SOLO coding agent to TRAE Work, signaling a broader focus on workflow collaboration. This convergence marks a pivotal industry consensus. The initial exploration phase, where companies rapidly built numerous overlapping agents for different scenarios, proved costly and inefficient. With open-source tools eroding technical barriers, competition has shifted from agent creation to resource consolidation and cost control. Historically, platform wars are won not by creating more products, but by simplifying them—as seen with browsers unifying web access and super-apps consolidating services. Now, the "super workbench" aims to become the unified AI entry point for work. This reflects a deeper market realization: the primary audience for AI is no longer just programmers (a market in the tens of millions) but all knowledge workers (a market of billions). The real opportunity lies in augmenting everyday tasks—managing emails, documents, data, and meetings—across the entire workday. The core battleground is becoming control over the primary AI entry point that employees use daily. Tencent's WorkBuddy leverages WeChat and Tencent Docs; Alibaba's Qianwen Office taps into DingTalk's organizational data; ByteDance's TRAE Work integrates with Feishu's workflows. Whoever owns this "super workbench" gains strategic control over orchestrating enterprise data and APIs. This shift is redefining enterprise software. Traditional SaaS applications, valued for their user interfaces, will recede into the background. Their core functionalities will be exposed as standardized "Skills" or APIs for the super workbench's agents to invoke. Software value will shift from selling user seats to charging based on API calls and outcomes delivered. The evolution of agents is moving through clear stages: first as novel standalone products, then as consolidated primary work entry points, and finally as pervasive, invisible capabilities embedded into the digital fabric. The recent moves by major tech firms signal the transition from the first stage into the second, accelerating toward the third. In the end, the most successful agent technology may become invisible—like electricity or the HTTP protocol—a fundamental, unnamed infrastructure powering work itself.

marsbit2m ago

Agent Race Ends, Super Workbench Takes Over

marsbit2m ago

Michael Saylor: 110 Reasons to Oppose BIP-110

Michael Saylor presents 110 arguments against Bitcoin Improvement Proposal (BIP) 110, a soft fork aimed at restricting certain non-monetary data storage uses (like inscriptions) on the Bitcoin blockchain. He acknowledges the proponents' valid concerns—such as node costs, fee pressure, and preserving Bitcoin's monetary focus—but fundamentally disagrees with the proposed solution. Saylor argues that BIP 110 represents a dangerous precedent of using consensus rules to enforce value judgments on transaction validity, moving away from Bitcoin's core principles of neutrality and permissionless innovation. His key objections are organized into eleven categories: 1) It violates neutrality and hard consensus by banning currently valid transactions. 2) It fails to meet the high burden of proof required for a consensus change, lacking concrete data on the alleged crisis. 3) Its seven bundled technical restrictions are overly broad, targeting generic script functionalities and blocking future upgrade paths. 4) It sacrifices compatibility and future optionality by closing off designed upgrade hooks. 5) Its temporary rules add significant complexity (grandfathering, expiry states) without sufficient justification. 6) The economic and security impacts, particularly on miner revenue and fee markets, are uncertain and unmodeled. 7) Superior, market-based tools (fee markets, relay/mining policies) already exist to manage blockchain load. 8) It stifles innovation by creating a chilling effect for developers. 9) Its modified activation mechanism (55% threshold, forced signaling) is aggressive and risks network splits. 10) The precedent it sets—using consensus to suppress disliked but legal uses—is more dangerous than the problem it aims to solve. 11) A better path exists: improving measurements, refining resource-based policies, and allowing market forces to work. Saylor concludes that Bitcoin's strength lies in its neutral rules, open markets, and hard consensus. Changing these foundational elements to target specific use cases is an unnecessary and risky "iatrogenic" intervention. He advocates for guarding Bitcoin's neutrality rather than acting as its redeemer.

marsbit18m ago

Michael Saylor: 110 Reasons to Oppose BIP-110

marsbit18m ago

Trading

Spot
活动图片