Vulnerability high
Severity
high
no CVSS

Google Patches Antigravity IDE Prompt Injection Sandbox Escape

Google Patches Antigravity IDE Prompt Injection Sandbox Escape is a high-severity remote code execution issue in Google Antigravity IDE, an AI agent. Affected: Google Antigravity IDE (pre-Feb 28 2026 patch).

Is Google Antigravity IDE running on your developers' machines? Anomity finds every device with this AI agent installed, shows the version, and can block it by policy. Detect this across your developers' devices with Anomity →

Key facts

ID
google-patches-antigravity-ide-prompt-injection-sandbox-escape
Severity
high
Attack type
Remote code execution, Prompt injection, Sandbox escape
Component
Google Antigravity IDE · AI agent
Known exploited
No known exploitation
Published
Last verified

What is this vulnerability?

Pillar Security researchers discovered a prompt injection vulnerability in Google's Antigravity agentic IDE that allowed arbitrary code execution by escaping the AI sandbox. By injecting the -X (exec-batch) flag through the find_by_name tool's Pattern parameter, an attacker could force fd to execute arbitrary binaries against workspace files. Combined with Antigravity's file-creation capabilities, this enables a full attack chain bypassing Strict Mode without user interaction once the prompt injection lands.

Affected versions

AffectedFixed in
Google Antigravity IDE (pre-Feb 28 2026 patch)No fixed version listed

Am I affected by this vulnerability?

You are affected if you run Google Antigravity IDE (pre-Feb 28 2026 patch).

How do I fix it?

Update Google Antigravity to latest version. Restrict AI agent file system access to project directories only. Audit IDE agent tool permissions. Avoid processing untrusted code or prompts in agentic IDE environments.

How do I find every machine running Google Antigravity IDE?

Anomity inventories the AI agents, MCP servers, IDE extensions and skills on each developer device, so you can see which machines run Google Antigravity IDE and which version, then enforce a policy on it.

References