Skip to content
Submit incident
AII-1000-32469
Well documentedThree or more independent publishers report the core facts.

Cursor AI code editor 0-day allows code execution from poisoned repositories

Cursor AI's code editor contained an unpatched 0-day vulnerability allowing arbitrary code execution from poisoned repositories on Windows. The vendor reportedly ignored the disclosure for five months before making a vague public acknowledgment.

15 December 2025 Β· Cursor Β· Security and cyberattacks
15 December 2025
AII-1000-324695 sources ↓
Quick facts
Category
Security and cyberattacks
Country
United States
Source
Dark Reading
Date
15 December 2025
Drafted by
Raidu Incident Detection

What happened

Mindgard disclosed a 0-day vulnerability in Cursor where a poisoned repository could trigger arbitrary code execution on Windows systems. The flaw exists in how Cursor resolves Git binaries when a project loads: Cursor checks several locations for git, including the workspace itself, and if a file called git.exe is placed at the root of a repository, Cursor executes it automatically when the project opens.

Mindgard first reported the bug on December 15, 2025, but Cursor did not respond. After a public LinkedIn post, Cursor's CISO replied privately, blaming a broken automation for the missed HackerOne invite. The initial HackerOne report was closed as 'informative and out of scope,' then reopened after Mindgard pushed back. Despite requests for updates in February, March, and April, Cursor provided no response for five months.

Reported impact

Cursor is used by more than half of Fortune 500 companies. The vulnerability exposes developers to arbitrary code execution attacks when opening repositories.

Index analysis

The issue exemplifies risks in AI-powered development tools that are designed to automate tasks and execute commands, including writing code, executing terminal commands, modifying files, and interacting with version control. Sandboxing an AI agent is fundamentally harder than sandboxing a static application because the agent is supposed to interact with the system in complex, context-dependent ways.

Governance controls

Cursor failed to respond to responsible disclosure over five months. An unnamed Cursor spokesperson later told Dark Reading that 'I can confirm we are addressing this and will get back to Mindgard accordingly,' but provided no timeline or confidence-instilling detail.

Classification

Category
Security and cyberattacks
Industry
Digital security
Country
United States
Harm status
Potential harm
Intent
Unknown
Timing
Post-deployment
Harmed
Business
Failure modes
Unsafe autonomous behaviour
Consequences
None recorded
AI systems
Cursor

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Dark Reading
Cursor IDE Auto-Executes Malicious Code in Poisoned Repos
14 July 2026
Crypto Briefing
Unpatched Cursor vulnerability exposes users to code execution risk
15 July 2026
It's FOSS
Cursor AI's Silence on a Critical Flaw Jeopardizes Millions of Users
15 July 2026
IT Security News - cybersecurity, infosecurity news
Critical Cursor 0-Day Flaw Allows Malicious Git Repos to Trigger Automatic Windows Code Execution - IT Security News
15 July 2026
International Business Times, Singapore Edition
Cursor Vulnerability in Windows Could Execute Code When You Open a GitHub Repository
16 July 2026