Anthropic Claude Finds Coldcard Wallet Flaw in Eight Minutes Amid $100M Bitcoin Hack

Summary

• Anthropic Claude identified the Coldcard wallet vulnerability within eight minutes, highlighting AI’s growing role in security code analysis efforts.
• Researchers linked the flaw to weak cryptographic randomness, enabling attackers to steal over $100 million worth of Bitcoin.
• Investigators detected another coordinated attack wave, prompting security experts to urge affected users to transfer funds immediately.

 

Security researchers revealed that Anthropic’s Claude independently identified the vulnerability behind the Coldcard wallet exploit after roughly eight minutes of code analysis. The reported discovery has become part of the investigation into a Bitcoin hardware wallet compromise that community estimates now place above $100 million.

According to researcher Medusa, Anthropic’s Claude analyzed the Coldcard wallet’s source code after receiving a single prompt requesting a security review. In a post on X, Medusa stated that the AI identified the same vulnerability exploited during the attack within about eight minutes.

According to Medusa, Claude located the flaw without prior knowledge of the exploit or additional instructions. The researcher noted that the result demonstrated how advanced AI systems can identify critical software vulnerabilities through automated code analysis.

Developers traced the vulnerability to Coldcard’s cryptographic random number generation. Hardware wallets depend on unpredictable random values when creating private keys and producing digital signatures. Consequently, weaknesses in that process can expose users to unauthorized access and financial losses.

Also Read: Bitcoin (BTC) Leads Broad Crypto Market Decline as Select Altcoins Defy the Selling Pressure

Researchers warn more Coldcard attacks may follow

Anthropic Claude’s reported discovery emerged as investigators continued tracking activity linked to the exploit. Moreover, the finding has renewed discussion about artificial intelligence assisting security researchers in detecting vulnerabilities before attackers exploit them.

As previously reported by 36Crypto, BlockTower Capital founder Ari Paul argued that the Coldcard compromise exposed a broader weakness affecting the cryptocurrency wallet industry. He explained that the incident reflected risks surrounding cryptographic randomness rather than a failure unique to a single manufacturer.

Meanwhile, researchers warned that the attack campaign may still be active, with Alex Thorn, Galaxy’s head of research, stating on X that Bitcoin blocks 960,778 through 960,792 contained 218 suspicious transactions affecting 462 wallet addresses. Those transactions reportedly transferred nearly 389 BTC during another coordinated wave of activity.

Furthermore, Thorn urged Coldcard users to move any remaining funds from affected devices immediately. He also recommended paying higher transaction fees to improve confirmation priority before attackers could interfere with pending transfers.

Conclusion

Anthropic Claude’s reported discovery has added another layer to the Coldcard investigation while highlighting AI’s growing role in software security. Meanwhile, researchers remain focused on monitoring suspicious blockchain activity as developers work to address the vulnerability linked to losses exceeding $100 million.

Also Read: Wheels Up Experience (UP) Stock Price Prediction 2026–2030: Can UP Rebound to $10?

The post Anthropic Claude Finds Coldcard Wallet Flaw in Eight Minutes Amid $100M Bitcoin Hack appeared first on 36Crypto.