Many students and new developers now use AI coding tools every day. Last week, one of those tools was turned into a trap. A tampered copy of Anthropic's Claude Code extension was offered as an update in the Cursor code editor. It tried to open a hidden back door into developers' computers.
The attack happened on Wednesday, 29 October 2026. Security companies kept publishing details over the next few days, and Anthropic clarified some points on 30 October. This post explains what happened, why it matters to you, and what to check on your own laptop.
What happened
Cursor is a popular AI code editor built on VS Code, and it has its own extension marketplace. On 29 October, Cursor said a modified version of Anthropic's Claude Code extension had been available in that marketplace from about 2:27 am to 1:35 pm Pacific Time. That is roughly 2:57 pm on 29 October to 2:05 am on 30 October, India time. BleepingComputer described it as a window of about 11 hours.
The bad builds were versions 2.1.312 and 2.1.313 of the extension anthropic.claude-code, for Windows and macOS. They used the real extension's name and ID, so to users they looked like a normal update. Cursor estimates that at most 5,000 users were affected.
Anthropic did not build or publish these versions. According to Anthropic, the attacker used access to Cursor's marketplace to publish the tampered copy. Cursor has not fully explained in public how the attacker got that access.
What the malicious code tried to do
According to Cursor and security researchers, the hidden code tried to:
- add the attacker's SSH key, labelled
svc-telemetry, to the user'sauthorized_keysfile, so the attacker could log in later; - turn on the computer's SSH service and open a reverse SSH tunnel to the attacker's server;
- set up persistence, meaning scripts that keep running even after a restart;
- send details about the machine to the attacker, including the username, hostname, public IP address, Git email address, and the names of environment variables linked to secret credentials.
There was some good news. Cursor said errors in the malicious code meant the SSH backdoor step would fail on most machines. Turning on SSH needs admin rights, and the extension usually did not have them. But the machine details were still sent to the attacker, and Cursor warned that uninstalling the extension may not be enough, because the malware leaves separate scripts behind.
Early reports said npm, Open VSX and Microsoft's VS Code Marketplace were also affected. That turned out to be wrong. Cursor was the only place that distributed the bad versions, and Microsoft confirmed they never appeared in its marketplace or on npm. In response, Cursor rotated its keys, checked other extensions for tampering, and temporarily took its marketplace offline to rebuild it.
Why it matters for learners and developers in India
This is a supply chain attack. The attacker did not trick users into installing an unknown app. They hid malware inside a trusted tool, delivered through a trusted channel, as a normal update. Even careful developers would not have noticed.
This matters a lot in India, where many students and freshers use AI editors such as Cursor, VS Code with AI extensions, and similar tools. Your laptop holds valuable things: GitHub access, SSH keys, cloud accounts, API keys in .env files, and sometimes access to your employer's or internship company's systems. If an attacker gets into a developer's machine, they can often reach many other systems from there.
There is also a career lesson. Companies now ask about secure development, even for junior roles. If you can explain an incident like this and describe how to protect against it, you will stand out in interviews.
What to do next
1. Check your own machine
- If you use Cursor and have the Claude Code extension, check whether version
2.1.312or2.1.313was installed or updated during the exposure window. - Open your SSH
authorized_keysfile (on macOS,~/.ssh/authorized_keys) and look for any key you do not recognise, especially one labelledsvc-telemetry. - If you were affected, follow Cursor's official security notice and its cleanup steps. Then change any passwords, tokens and API keys that were stored on that machine. Security firm Sophos also recommends rotating credentials.
2. Build safer habits
- Install fewer extensions. Every extension can run code on your computer. Remove the ones you do not use.
- Do not run your editor as administrator. In this attack, the lack of admin rights stopped the worst damage.
- Keep secrets out of code. Never commit
.envfiles. Use short-lived tokens with limited permissions where possible. - Protect your accounts. Turn on two-factor authentication for GitHub, email and cloud accounts.
3. Learn the basics behind the news
To really understand attacks like this, learn how SSH keys work, what authorized_keys does, and how Git uses your identity and credentials. Our free Git tutorial is a good start. To understand processes, networking and operating systems more deeply, try our free Computer Science tutorial.
AI tools make you faster, but you are still responsible for what runs on your machine. Trust your tools, but check them too.
Keep using AI coding assistants. They are valuable skills for your career. Just treat every extension and update as code that runs with your permissions, because that is exactly what it is.
Sources
- Cursor: Security notice — malicious version of Anthropic.claude-code (2.1.312 and 2.1.313) identified in the Cursor VS Code Extension Marketplace
- BleepingComputer: Malicious Claude Code extension on Cursor planted SSH backdoors
- The Hacker News: Official Claude Code Extension on Cursor Marketplace Trojanized with Persistent SSH Backdoor
- Sophos: Malicious Claude Code extension distributed through Cursor marketplace
- Dark Reading: Devs Hit by Trojanized Claude Code Extension for Cursor
Written by
SukhNexus Academy
SukhNexus IT Solutions, Bathinda