HIGHmalware·

Cursor IDE Supply Chain Compromise: Malicious Claude Code Extension

Cursor, an AI-native code editor, confirmed a security breach that led to the distribution of a malicious ‘Claude Code extension’ through its marketplace. Attackers gained access by stealing credentials from AI infrastructure provider Fireworks AI, enabling them to compromise Cursor’s build and distribution pipeline.

This report was researched and drafted by an AI agent and reviewed by a human analyst prior to publication. View the agent workflow →

Overview

Cursor, an AI-native code editor, experienced a security breach that resulted in the distribution of a malicious ‘Claude Code extension’ via its official marketplace. The compromise stemmed from attackers stealing credentials belonging to Fireworks AI, an infrastructure provider used by Cursor, allowing them to inject malicious code into the extension’s build process. This incident represents a supply chain attack, potentially exposing users who installed the extension to arbitrary code execution or data theft.

Technical Analysis

Attackers gained unauthorized access to Cursor’s systems by compromising credentials associated with Fireworks AI, a third-party AI infrastructure provider. This access allowed them to inject malicious code into the ‘Claude Code extension’ distributed through the Cursor marketplace. The specific payload and capabilities of the malicious extension are not detailed in the source, but such extensions typically aim to exfiltrate sensitive data (e.g., source code, credentials, API keys) or establish persistence within the developer’s environment. The attack vector was the official Cursor marketplace, making it a supply chain compromise where trusted software was weaponized.

Detection

  • Extension Inventory: Check installed Cursor extensions for ‘Claude Code extension’. Immediately remove if found.
  • Process Monitoring: Monitor cursor.exe (or equivalent IDE process on Linux/macOS) for unusual child process creation (e.g., cmd.exe, powershell.exe, sh, python, curl, git commands with suspicious arguments).
  • Network Monitoring: Look for outbound connections from cursor.exe to unusual or unapproved external IP addresses or domains, especially those not related to Cursor’s legitimate update or AI services.
  • File System Monitoring: Monitor for suspicious file writes or reads by cursor.exe in sensitive directories (e.g., .ssh, .aws, .git, temporary directories).

Sigma Detection Rules

⚠️ AI-generated detection rules. These are experimental starting points. Review field names, EventIDs, and logic against your environment’s schema before deploying. Tune to reduce false positives.

Suspicious Child Process from Cursor IDE

title: Suspicious Child Process from Cursor IDE
id: 9a7b6c5d-4e3f-2a1b-0c9d-8e7f6a5b4c3d
status: experimental
description: Detects suspicious child processes spawned by the Cursor IDE process, which could indicate compromise via a malicious extension.
logsource:
  product: windows
  service: sysmon
detection:
  selection:
    ParentImage|endswith: '\\cursor.exe'
    Image|endswith:
      - '\\cmd.exe'
      - '\\powershell.exe'
      - '\\pwsh.exe'
      - '\\bash.exe'
      - '\\sh.exe'
      - '\\curl.exe'
      - '\\wget.exe'
      - '\\python.exe'
      - '\\perl.exe'
      - '\\ruby.exe'
      - '\\node.exe'
  condition: selection
level: high

Mitigations

  1. Remove Malicious Extension: Immediately uninstall the ‘Claude Code extension’ from Cursor IDE if present.
  2. Credential Rotation: Rotate all credentials, API keys, and tokens used within the Cursor IDE or stored in its workspace, especially if the extension was installed.
  3. Review Permissions: Audit permissions granted to Cursor IDE and its extensions. Restrict network access and file system access where possible.
  4. Supply Chain Vigilance: Implement stricter vetting for third-party integrations and regularly audit dependencies and extensions for integrity.
  5. Endpoint Detection & Response (EDR): Ensure EDR solutions are actively monitoring IDE processes for anomalous behavior.

References

  • https://www.bleepingcomputer.com/news/security/cursor-confirms-security-breach-led-to-malicious-claude-code-extension/

Indicators of Compromise

No public IOCs available at time of writing.

MITRE ATT&CK

  • T1195.002 — Compromise Software Supply Chain
  • T1059 — Command and Scripting Interpreter
  • T1552 — Unsecured Credentials
🤖 AI Attribution
Generated by gemini-2.5-flash ·
1,487 input / 1,116 output tokens ·
Reviewed and approved by a human analyst before publication
#malware#credential-theft#extension#high#ide#linux#macos#malware#supply-chain#windows