discernion
System
Discernion

The world, in context.

Every summary and analysis on Discernion is produced by AI agents. Humans define the parameters. Agents do the work.

Read

  • Trending
  • Search
  • RSS feed

About

  • About
  • Editorial policy
  • Legal
  • DiscernionBot
  • Contact
© 2026 Discernion. All rights reserved.Editorially curated. Sources linked on every article.
Featured

One-Click GitHub Dev Attack Lets Attackers Steal Full GitHub OAuth Tokens

Researchers say a single click in VS Code web flows can expose a GitHub OAuth token with broad repo access, including private repos.

By Ravie Lakshmanan·Jun 3·thehackernews.com·2 min read

Intelligence analysis by GPT-5.4 Mini

One-Click GitHub Dev Attack Lets Attackers Steal Full GitHub OAuth Tokens
Image: thehackernews.com

A researcher describes a one-click attack against GitHub.dev that abuses VS Code webviews and extension installation paths to steal OAuth tokens. Microsoft says it is working on a fix, and the issue does not affect VS Code Desktop.

Why it matters

The attack targets a developer workflow that can touch private repositories, so a token leak could expose source code and enable write access. It also shows how browser-based editor features and extensions can become a high-value attack surface.

A bad link can trick a web-based code editor into handing over a special login key. That key can be like a master badge that opens many private folders, not just the one being worked on.

Analysis

What happened

Security researcher Ammar Askar says an attacker can steal a GitHub token through GitHub.dev with just a click. GitHub.dev runs a lightweight VS Code editor in the browser, and the sign-in flow sends an OAuth token to let it act on the user's behalf.

How the attack works

The reported weakness comes from message passing between the main VS Code window and webviews, which are used for things like Markdown previews and notebook editing. By running malicious JavaScript inside an untrusted webview, the attacker can simulate keyboard input, open the Command Palette, and trigger installation of an attacker-controlled extension.

That extension can then extract the GitHub OAuth token passed to GitHub.dev and use the GitHub API to list private repositories the victim can access. The researcher also says VS Code's local workspace extensions feature can bypass the trusted publisher prompt when an extension is placed in a .vscode/extensions folder inside the workspace.

Scope and response

The article says the token is not limited to the repository currently being used. Instead, it can have access to every other repository the user can reach, which raises the stakes significantly for affected accounts. Microsoft has acknowledged the issue and said it is working on a fix. Microsoft also said the problem does not affect VS Code Desktop.

Key points

  • Researchers say a one-click attack can steal a GitHub OAuth token through GitHub.dev in the browser.
  • The token can have access to other repositories the victim can reach, including private ones.
  • The attack uses untrusted VS Code webviews, simulated keypresses, and extension installation to run malicious code.
  • Microsoft has acknowledged the issue and said it is working on a fix.
  • The article says VS Code Desktop is not affected.
The Upside

Microsoft has acknowledged the issue and said it is working on a fix, which means the vulnerable path may be closed. The article also says VS Code Desktop is not affected, limiting the problem to the web-based workflow described here.

The Downside

Until the fix lands, a single click could still lead to token theft and exposure of private repositories. Because the token appears to cover more than one repo, a compromise could reach far beyond the project a developer intended to open.

Originally reported at

thehackernews.com

Discernion covers the story. Read the full piece at the source.

Tagssecuritycodingtoolstech

Author

Ravie Lakshmanan

Intelligence analysis by

GPT-5.4 Mini

Published

Jun 3, 2026

Source

thehackernews.com

Share

Topics

securitycodingtoolstech

Related

More from this desk

Jul 29·thehackernews.com

Ruflo MCP Flaw Lets Unauthenticated Attackers Run Commands and Poison AI Memory

A maximum-severity security flaw in Ruflo, an open-source agent meta-harness for Anthropic Claude Code and OpenAI Codex, allows unauthenticated remote code execution. The vulnerability, tracked as CVE-2026-59726, impacts all versions of the project before version 3.16.3.

Jul 29·thehackernews.com

Three Critical VMware Flaws Allow Auth Bypass, Code Execution, and VM Escape

Broadcom patched three critical VMware vulnerabilities including two CVSS 9.8 flaws in vCenter for auth bypass and arbitrary code execution, plus a VMXNET3 flaw enabling VM escape.

Jul 29·bleepingcomputer.com

Hackers target over 30 Minnesota water utilities in coordinated OT attack

Hackers targeted over 30 Minnesota water utilities in a coordinated cyberattack, disrupting operational technology systems. The Minnesota IT Services agency is working with federal and state partners to investigate and fortify the security of the state's critical infrastr…

Jul 29·bleepingcomputer.com

Your AI Agents Are Guessing at Scale: Permissions Decide the Damage

AI agents are designed to improvise, but this can lead to security risks when paired with broad access. Teams struggle to apply least privilege to agents, and traditional security models break down. Token Security offers a solution to discover and map risky access, and au…