Antigravity adds security notice warning users about AI coding agent risks

TestingCatalog reports that Antigravity is showing a new notice warning that AI coding agents have security limitations, including potential data exfiltration and code execution risks.

The notice tells users to avoid processing highly sensitive data and to verify all actions the agent takes. The report is based on a screenshot-style post from the TestingCatalog account on X and does not include an official Antigravity statement.

The same post suggests a new agent may launch next week, according to TestingCatalog.