On this page (9 sections)
A critical vulnerability, CVE-2026-120779, was published on 2026-11-10 for Visual Studio Code. The NVD record describes it as improper control of generation of code ('code injection') that allows an unauthorized attacker to execute code over a network. It carries a CVSS score of 9.6.
If you or your team write, review or deploy website code in VS Code, this matters: developer machines often hold SSH keys, cloud credentials, API tokens and deploy access. Public detail is very limited at the moment, so this article separates what is confirmed from what is not.
At a glance
What is confirmed so far
The only primary source available is the NVD entry, published 2026-11-10. It gives a single-sentence description, the CWE-94 classification and the 9.6 score. The record points to Microsoft's Security Update Guide entry as its reference.
Only one independent publisher covers this story so far, and the CVE record shows no CISA Known Exploited Vulnerabilities listing. That is not proof of safety, only a statement about what is publicly known today.
What is not yet known
- Which Visual Studio Code versions are affected and which release fixes the flaw.
- The exact attack vector, for example whether a user must open a file, folder or link, or whether the editor must be listening on a network port. The description says only that the attack works over a network by an unauthorized attacker.
- Whether a full CVSS vector string (attack complexity, user interaction, scope) exists. Only the 9.6 score is available to us.
- Whether the flaw affects remote or web variants of the editor, or only the desktop application.
- Whether any exploitation or proof of concept exists in the wild.
Why code injection in an editor is serious
CWE-94 covers cases where software builds code from input that an attacker can influence without properly restricting it, so the attacker's data ends up being run as code. In a code editor, that code would run with the privileges of the logged-in developer. For many teams that means access to source repositories, password managers' unlocked sessions, cloud CLI credentials, SSH agents and deployment keys.
A compromised developer workstation is a common stepping stone to compromising the websites and servers that developer manages. That is why a critical editor flaw deserves attention even though it is not a web server bug.
Who is realistically exposed
- Developers and agency staff using Visual Studio Code on laptops or desktops.
- Shared build machines, jump hosts or remote development environments where VS Code or its server components are installed.
- Teams that open untrusted repositories, archives or links in the editor. Whether this is relevant to this CVE is unconfirmed, but it is a sensible area to be careful with until details are published.
How to check your exposure
First, find out which version each machine runs. This read-only command prints the installed Visual Studio Code version, commit and architecture. Run it in a terminal on Windows, macOS or Linux if the code command is on your PATH:
code --versionThis read-only command lists installed extensions with their versions, which helps you review what is running inside the editor:
code --list-extensions --show-versionsCompare the version against the fixed version named in the Microsoft advisory for CVE-2026-120779 once it lists one.
What to do now
- Open the Microsoft advisory for CVE-2026-120779 and note the affected and fixed versions.
- Inventory every machine and server that has Visual Studio Code installed, including build hosts and remote development hosts.
- Update Visual Studio Code to the latest release through its built-in updater or your software management tool, and confirm with code --version afterwards.
- If your organization pins or delays editor updates, prioritize this one for machines with production, cloud or repository admin access.
- Until you know the attack vector, avoid opening untrusted projects, files or links in VS Code on sensitive machines.
- Review and remove extensions you do not need, and keep the rest updated.
- Make sure developer machines use separate, least-privilege credentials and that production keys are not stored in plain text in project folders.
- Rotate credentials from any machine you suspect was compromised, and follow our guide on what to do if a website is hacked.
Detection and monitoring
No indicators of compromise have been published in the sources we have. In the meantime, watch for unexpected child processes launched by the editor, unfamiliar outbound connections from developer machines, and new SSH keys or access tokens appearing in your repositories and cloud accounts.
For the sites you manage, regular outside-in checks help catch the consequences of a stolen credential, such as unexpected file changes or new scripts. A read-only scan or change monitoring can surface those symptoms, but neither can see inside a developer laptop, so endpoint updates remain your main defense. Our guide to checking whether a website is hacked lists signs to look for.
Timeline
| Date | Event |
|---|---|
| 2026-11-10 | CVE-2026-120779 published in NVD with CVSS 9.6, CWE-94 |
| 2026-11-10 | Microsoft Security Update Guide entry referenced by the CVE record |
This article will be most useful once the Microsoft advisory details are confirmed. Recheck the advisory for affected builds, the fixed release and any mitigations before closing this out internally.
Sources
This article is based on the public sources below. Details can change as vendors and researchers publish more; check the original advisories before acting on version numbers.
