Cursor’s November 13, 2026 update traces its marketplace malware incident to an exposed vendor key and credentials in stored coding sessions.
An attacker used a Cursor-issued key exposed by Fireworks AI to copy a read-only bucket of agent traces, mostly from 2025, from sessions outside Privacy Mode. Cursor says those traces were permitted for model training. Unredacted credentials led into selected internal SaaS tools, exposing archived non-production repositories, limited support records and further credentials.
A stolen internal credential then let the attacker place unauthorized Claude Code copies directly in Cursor’s extension marketplace cache on October 29. Cursor says at most 5,000 users potentially received them. Open VSX, the upstream marketplace source, never carried the malware, according to Cursor.
Microsoft Security Research’s October 30 analysis said Anthropic confirmed its systems were not compromised and it had not developed the malicious packages. Microsoft also confirmed that the packages were never distributed through its Visual Studio Marketplace or the official Claude Code npm package.
Cursor says Privacy Mode prompts and code, customer repositories and official application updates were unaffected. It believes the incident is contained; the last attacker activity it observed, on November 1, was blocked.
Fireworks describes the vendor-side timeline
In its November 9 bulletin, Fireworks said a third-party cloud provider hosting its research-and-development machines suffered a security incident on October 23. An unauthorized party reached Fireworks development machines on October 24 and ran an unauthorized workload, causing additional customer data exposure.
In a November 17 update, Fireworks said its investigation had identified two separate incidents and that both were contained and remediated. It said the second ran from October 23 to October 30 and exposed account and billing information for accounts created before October 30, plus datasets, models or secrets for a small number of customers. Payment card numbers were not involved. Unit 42 completed its investigation of the first incident; AprosSyber is investigating the second. Those findings concern Fireworks’ environment rather than the later activity described by Cursor.
Why removing the extension may not end the risk
Cursor’s original security notice identifies tampered Anthropic.claude-code versions 2.1.312 and 2.1.313 for Windows and macOS. They were available between about 2:27 a.m. and 1:35 p.m. Pacific time on October 29, 2026.
Microsoft’s code analysis describes attempts to collect system information, add an attacker-controlled Secure Shell (SSH) key and establish a persistent remote-access tunnel. Successful access depended on host privileges and SSH configuration, but keys and persistence artifacts could still be written when tunnel access failed.
Microsoft recommends preserving evidence and isolating devices showing evidence of malicious code execution or backdoor installation. Its analysis warns that removing or updating the extension does not remove artifacts written outside the extension’s installation directory.
Cursor’s notice directs affected customers to check their machines, remove unauthorized keys and persistence mechanisms, and assess whether rebuilding is necessary. It also calls for rotating credentials and sessions accessible from affected hosts, using a trusted system.

Tech Help Canada Staff researches, writes, and reviews practical content for business owners and professionals. Our coverage spans business, marketing, SEO, technology, and the tools and systems people use to grow and operate online. We focus on clear, useful information backed by research, hands-on experience, and editorial review. Learn more about our team and editorial standards. Need help with something? Contact Us







