A new security report claims that Anthropic was hacked in an operation designed to deliver a remote access trojan to Claude developer laptops across the globe. According to the account published by DoublePulsar, the attack reached machines used by people building with Claude, turning a trusted AI development workflow into a distribution channel for malicious software. The details point to a serious compromise that could affect developers well beyond a single organization.
The reported attack matters because Claude Code and similar AI coding assistants run with broad access on developer machines. A remote access trojan in that position can read source code, capture credentials and move laterally into corporate networks. Developers often install extensions and tooling with minimal scrutiny, which makes the AI toolchain an attractive target for attackers seeking a quiet foothold.
The disclosure arrives alongside separate reporting about backdoors found in official Claude Code extensions, adding to concerns about the integrity of the plugin ecosystem around AI coding tools. Taken together, the two stories suggest that the security model for AI developer tooling has not kept pace with its rapid adoption. Organizations deploying these tools may need to audit extensions and monitor developer endpoints more closely.
Anthropic has not issued a detailed public statement in the material reviewed here, and the full scope of the compromise remains unclear. What is already evident is that the incident will intensify scrutiny of how AI vendors secure their distribution channels. For developers, the practical lesson is to treat AI coding tools and their extensions as high value attack surfaces rather than convenience utilities.