A reported compromise of the official Claude Code extension on the Open VSX registry has raised a serious software supply-chain concern for businesses using AI-assisted development tools. According to cybersecurity researchers cited by The Hacker News, the malicious version was designed to install a persistent SSH backdoor on affected systems. The issue was reportedly traced to a compromise of the Cursor Marketplace, which Cursor briefly took offline to remove the malware.

What happened

The reported incident involves the official Claude Code extension distributed through Open VSX, an extension registry used by some code editors and development environments. Researchers said the compromised extension attempted to install a persistent SSH backdoor, which could allow continued remote access to an infected host.

Based on the source report, the incident was later linked to a compromise affecting the Cursor Marketplace. Cursor reportedly took the marketplace down temporarily while removing the malicious content. At the time reflected in the source, the publicly reported facts indicate a supply-chain style attack affecting a trusted software component rather than a user downloading an obviously suspicious tool.

Why it matters for European businesses

For European businesses, this is not just a developer-tool incident. Many companies are now introducing AI coding assistants, code editors, extensions and agent-based development workflows into day-to-day software operations. When a trusted extension or marketplace is compromised, the risk can spread into company laptops, development servers, internal repositories and cloud environments.

An SSH backdoor is particularly serious because it may enable persistent remote access. In a business context, that can create exposure across source code, deployment pipelines, credentials, internal documentation and connected infrastructure. Even organisations with relatively small development teams can be affected if developers have broad access to production systems, cloud accounts or customer data.

The case also matters because AI tooling is increasingly being adopted outside large software companies. SMEs, digital agencies, e-commerce businesses and in-house IT teams may rely on packaged extensions and managed marketplaces rather than running tightly controlled internal development environments. That convenience can accelerate productivity, but it also increases dependence on the security of upstream vendors and distribution channels.

Who may be affected

The immediate risk is highest for organisations that use:

• AI-assisted coding tools such as Claude-related development extensions
• Open VSX or alternative extension registries in development workflows
• Cursor Marketplace or similar curated extension ecosystems
• Developer endpoints with SSH access to servers, cloud workloads or internal environments
• Small engineering teams where individual developers may hold privileged access across multiple systems

Businesses working with external software partners or agencies should also consider whether contractors and suppliers use the affected tools, as supply-chain exposure may extend beyond internal staff devices.

What companies should consider

Businesses do not need to stop using AI development tools, but they should review how those tools are introduced and monitored.

• Identify exposure: Check whether your teams installed the Claude Code extension via Open VSX or used the relevant marketplace during the reported incident window.
• Investigate developer endpoints: Review affected machines for unauthorised SSH keys, persistence mechanisms, unexpected outbound connections and changes to shell or startup configuration.
• Rotate credentials where appropriate: If there is any chance a compromised machine had access to sensitive infrastructure, review and rotate SSH keys, API tokens and other privileged credentials.
• Audit access paths: Confirm which developer devices can reach production systems, cloud consoles, repositories and CI/CD pipelines, and reduce unnecessary privileges.
• Tighten extension governance: Consider allow-listing approved extensions, restricting marketplace sources and documenting which registries are permitted in corporate environments.
• Strengthen monitoring: Ensure endpoint detection, central logging and alerting cover developer workstations as well as servers.
• Review supplier risk: If agencies or contractors contribute code or maintain infrastructure, ask what controls they apply to IDE extensions and AI tooling.

For regulated or security-sensitive businesses, the incident is a reminder that developer environments are part of the attack surface. AI adoption can improve productivity, but it should be paired with software supply-chain controls, endpoint security and clear rules for third-party tooling.

Companies should also watch for further statements from the affected vendors and researchers, as the exact scope, timeline and remediation guidance may evolve as the investigation continues.