Infostealer adds AI assistants and developer tools to its targets, Check Point Research reports
Check Point Research has analysed an information stealer that, beyond the usual targets, includes dedicated routines for AI applications such as Claude, ChatGPT Desktop, Codex, Cursor, Windsurf and GitHub Copilot. Developers and organisations running these tools on machines that hold cloud credentials are most exposed.
What to do now
Check Point's research team published on 23 October 2026 an analysis of an infostealer with a wider collection scope than usual. Alongside the classic targets of this malware family, namely browser-stored passwords, session cookies and cryptocurrency wallets, it carries routines built specifically to harvest data from AI applications and developer tooling.
The software named as targets includes Claude, ChatGPT Desktop, Codex, Cursor, Windsurf and GitHub Copilot. The malware also goes after cloud credentials and sensitive files stored in local development environments.
The researchers see this as a sign that attackers are paying closer attention to developer workstations. These machines tend to hold API tokens, access keys and AI assistant conversation histories that may include source code or internal information.
The available excerpt does not give the malware's name, how it is distributed, which operating systems it runs on, or any identified victims. It also lists no indicators of compromise; these may be included in Check Point Research's full report.
For small and mid-sized organisations, the main danger lies in what happens after the theft: a single cloud token or API key taken from one laptop can open access to code repositories, production services or pay-per-use accounts.
What to do now
- Read the full Check Point Research report to obtain any indicators of compromise and load them into your detection tools.
- Inventory machines running AI assistants and developer tools such as Cursor, Windsurf, Codex or GitHub Copilot, and confirm they are covered by your endpoint protection.
- As general good practice, rotate API tokens and cloud credentials stored on developer workstations if there is any doubt, and favour short-lived secrets.
- Watch for unusual sign-ins to cloud accounts and AI services used by your teams, which may indicate reuse of stolen secrets.