Skip to content
The Cybersecurity Focus
The Cybersecurity Focus

Cybersecurity news, threat intelligence, and vulnerability research.

The Cybersecurity Focus

Cybersecurity news, threat intelligence, and vulnerability research.

  • Home
  • Home
The Cybersecurity Focus

Cybersecurity news, threat intelligence, and vulnerability research.

The Cybersecurity Focus

Cybersecurity news, threat intelligence, and vulnerability research.

  • Home
  • Home
BreachexploitVulnerability

Cursor Security Breach Caused by Malicious Claude Code Extension

By ogwatermelon
November 15, 2026 4 Min Read
0

 

November 15, 2026

Cursor confirmed a security breach that let hackers distribute a malicious Claude Code extension through its marketplace after stealing credentials from AI infrastructure provider Fireworks AI. The incident, disclosed on November 14, 2026, exposed up to 5,000 developers to malware designed to install persistent SSH backdoors.

What Happened: Cursor Security Breach Delivers Malicious Claude Code Extension

Cursor published an incident report on Friday detailing how attackers exploited a stolen access key from Fireworks AI to compromise internal systems. The attackers used the exposed key to copy data from a read-only storage bucket that Cursor used to share information with the third-party AI provider.

This bucket contained historical AI coding-agent traces, mostly from 2025, collected from users who had not enabled Privacy Mode. Some of these traces held unredacted API credentials. The attackers leveraged those credentials to access additional Cursor systems, including archived source code repositories, customer support data, product usage metrics, and summaries of AI coding-agent conversations.

Furthermore, the attackers obtained a credential that allowed them to distribute extensions through Cursor’s VS Code extension marketplace. On October 29, they published malicious copies of Anthropic’s Claude Code extension directly on the cache used to serve the marketplace. The malicious versions, 2.1.312 and 2.1.313, remained available for approximately 11 hours before Cursor removed them.

Technical Details of the Cursor Claude Code Extension Attack

The attackers modified the legitimate Claude Code extension to execute malicious JavaScript that attempted to install persistent SSH backdoors on developers’ computers. The malware performed several actions:

  • Collected system information from the infected machine
  • Installed an attacker-controlled SSH key for unauthorized remote access
  • Attempted to enable SSH access if it was not already running
  • Configured persistence mechanisms to maintain remote access across reboots

However, security researchers found coding errors in the malware that likely prevented the backdoor from working on many Windows and macOS systems where SSH was not already enabled. Cursor estimates that up to 5,000 users who requested affected extension versions during the exposure window may have received the malicious builds.

Cursor also emphasized that the attack did not compromise Anthropic’s legitimate extension or Open VSX, the extension marketplace Cursor uses. In addition, production infrastructure, official application releases, customer code repositories, and data belonging to users with Privacy Mode enabled were not affected.

Business and Operational Impact

The Cursor security breach carries significant implications for developers, enterprises, and the broader AI development ecosystem:

  • Developer workstation compromise: Up to 5,000 developers may have installed malware designed to maintain persistent remote access
  • Supply chain risk: A compromised AI coding extension can poison downstream codebases, CI/CD pipelines, and container images
  • Credential exposure: Historical coding-agent traces contained unredacted API credentials, leading to further system compromise
  • Third-party dependency fragility: The breach originated at Fireworks AI, highlighting how vendor relationships can become attack vectors
  • Reputational damage: Cursor has moved all inference traffic away from Fireworks and rebuilt its extension marketplace

Moreover, the incident comes amid a wave of attacks targeting AI developer platforms that Wiz has dubbed “Black October.” These attacks targeted development infrastructure to steal authentication credentials, especially those providing access to AI models.

Mitigation and Recommendations

Cursor has taken several remediation steps, including revoking and rotating exposed credentials, rebuilding its extension marketplace, and redeploying production infrastructure. However, affected developers must also act.

Immediate Actions for Developers

  1. Check whether you installed or updated the Claude Code extension in Cursor on or around October 29, 2026
  2. Review installed extensions for versions 2.1.312 or 2.1.313 of the Claude Code extension
  3. Remove any suspicious Claude Code installations and reinstall from the official source
  4. Audit SSH authorized_keys files for unknown entries
  5. Rotate any API credentials that were shared with Cursor support or used in non-Privacy Mode sessions
  6. Review system logs for unauthorized SSH connections or unexpected JavaScript execution

Enterprise Security Guidance

Security teams should treat any developer workstation that installed the malicious extension as potentially compromised. Consequently, organizations should enforce extension whitelisting, require Privacy Mode for sensitive AI coding workflows, and segment developer environments from production networks. In addition, teams should monitor for signs of lateral movement from developer endpoints.

Bottom line: The Cursor security breach shows how a single stolen third-party credential can cascade into a supply chain attack against thousands of developers. Treat AI coding extensions as high-trust software, verify extension sources, and enable Privacy Mode to reduce exposure.

Incident Summary

Incident: Cursor Security Breach — Malicious Claude Code Extension
Affected Systems: Cursor VS Code extension marketplace; developers who installed Claude Code extension versions 2.1.312 or 2.1.313
Disclosure Date: November 14, 2026
Attack Date: October 29, 2026
Patch Status: Malicious extensions removed; marketplace rebuilt; credentials rotated
Exploitation Status: Confirmed compromise via stolen Fireworks AI credential; up to 5,000 developers exposed

References

  1. Cursor, “Security Incident Update,” November 14, 2026, https://cursor.com/blog/security-incident-update
  2. Cursor, “Security Notice: Claude Code Extension,” October 30, 2026, https://cursor.com/blog/security-notice-claude-code-extension
  3. Bill Toulas, “Cursor confirms security breach led to malicious Claude Code extension,” BleepingComputer, November 14, 2026, https://www.bleepingcomputer.com/news/security/cursor-confirms-security-breach-led-to-malicious-claude-code-extension/
  4. Wiz, “Black October for AI Developer Platforms,” November 2026, https://www.wiz.io/blog/black-october-for-ai-developer-platforms
  5. Lawrence Abrams, “Hackers compromise 4,800 GitHub accounts to steal cloud credentials,” BleepingComputer, November 12, 2026, https://www.bleepingcomputer.com/news/security/hackers-compromise-4-800-github-accounts-to-steal-cloud-credentials/

 

 

Tags:

BreachExploitVulnerability
Author

ogwatermelon

Follow Me
Other Articles
Previous

N-able 4EverAble2Shell Zero-Day Actively Exploited

Next

Mior and Skyfall Attacks Break AMD SEV-SNP Confidentiality

AI Android Backdoor Botnet Breach Cloud CVE Exploit Hack Incident Infostealer KEV Linux Mac Malware Network News Phishing Ransomware RCE supp supply chain Virtualization Vishing Vulnerability Windows World Worm Zero Day

Meta

  • Log in
  • Entries feed
  • Comments feed
  • WordPress.org

Archives

  • November 2026
  • October 2026
  • September 2026
  • August 2026
  • July 2026
  • June 2026
  • May 2026

Categories

  • AI
  • Android
  • Apple
  • Backdoor
  • BotNet
  • Breach
  • Cloud
  • CVE
  • exploit
  • Hack
  • Incident
  • Infostealer
  • KEV
  • Linux
  • Mac
  • Malware
  • News
  • Phishing
  • Ransomware
  • RCE
  • supply-chain
  • Uncategorized
  • Virtualization
  • Vishing
  • Vulnerabilities
  • Vulnerability
  • Windows
  • World
  • Worm
  • Zero Day
Copyright 2026 — The Cybersecurity Focus. All rights reserved.