Malicious Claude Code Extension Threatens Developer Workstations

Rogue VS Code Extension Deploys Backdoor

Software developers downloading coding tools from the Open VSX registry faced critical infrastructure exposure when version 2.1. 313 of the Claude Code extension deployed an active backdoor. According to security findings published October 30 by OpenSourceMalware researchers, the fraudulent package mirrored legitimate tools while executing hidden payloads upon initialization.

The malicious script initiated remote connections to external servers controlled by attackers. The backdoor maintained persistence across system reboots, meaning standard workspace closures failed to remove the intrusion vector.

It was a tampered version of the Claude Code extension which slipped into the Open VSX registry carrying a hidden backdoor. Researchers said the malicious code launched automatically when the extension loaded, attempted to connect to an attacker-controlled server, and was designed to survive system reboots, turning a routine developer tool into a persistence mechanism.

Blockchain Infrastructure and Private Keys Exposed

Crypto developers operating on infected machines confronted immediate risks regarding wallet credentials, private keys, and exchange API access. Security researchers noted that while the code lacked specific blockchain-targeting routines, compromised workstations housed critical validator and treasury credentials.

The breach threatened substantial digital asset reserves if malicious actors exfiltrated sensitive configuration files. Investigations confirmed the unauthorized modification affected the Open VSX repository specifically.

While not built specifically for crypto theft, the attack posed outsized risks for blockchain developers. Machines used to build DeFi protocols, wallets and trading infrastructure often store private keys, wallet credentials and exchange API keys. A successful compromise could have provided attackers with access to sensitive systems and high-value digital assets.

Anthropic Official Marketplace Remains Unaffected

Security audits confirmed that Anthropic distribution channels on official marketplaces sustained no structural compromises during the incident. OpenSourceMalware investigators could not determine how bad actors injected the malicious payload into the third-party registry.

Engineers running the affected version were advised to treat local machines as fully compromised. Immediate forensic audits superseded simple extension removal due to the persistent nature of the installed malware.

Researchers were careful to distinguish the rogue package from Anthropic’s legitimate distribution channels. The findings point to a compromised extension distributed through Open VSX, not the official Visual Studio Marketplace release. No evidence was presented showing Anthropic’s systems had been breached, narrowing the incident to a supply-chain compromise affecting a specific third-party repository. No doubt we'll be hearing about more of these stories in the months and years to come.