Latest / Frontier models and safety

IncidentFrontierInternationalConfirmed

Check Point discloses ChatGPT Lockdown Mode bypass, patched by OpenAI

See how this story connects

Check Point researchers disclosed a bypass of Lockdown Mode, the optional ChatGPT setting OpenAI introduced in June 2026 to limit data exfiltration from prompt injections. The flaw sat in HTML page previews inside code blocks, where a popup window could escape the nested sandbox and be used to exfiltrate chat data or deliver malicious payloads. OpenAI fixed the vulnerability on September 18, 2026, after responsible disclosure.

Why it matters

Users who relied on Lockdown Mode had a flaw that could enable silent chat data theft until OpenAI’s fix on 18 September 2026.

When
5 Nov 2026
ID
WTA-N-2026-0752
Who
OpenAI, Check Point; Eli Smadja
System
ChatGPT Lockdown Mode, HTML previews in code blocks
Operator
OpenAI
Type
a popup from a nested frame escaped the preview sandbox and got around Lockdown Mode, allowing chat data theft, phishing and malware delivery
Status
fixed by OpenAI on 18 September 2026; disclosed by Check Point, through The Hacker News · Incident ledger

Summary written with AI by wheresthe.ai, from The Hacker News. Checked against the source on 5 Nov 2026.

Source copy: saved on 5 Nov 2026 to check this story and not republished, because the text belongs to its publisher. Also held by the Internet Archive.

Cite this story

WTA-N-2026-0752: “Check Point discloses ChatGPT Lockdown Mode bypass, patched by OpenAI.” wheresthe.ai, 5 Nov 2026. https://wheresthe.ai/d/intl-check-point-discloses-chatgpt-lockdown-mode-bypass-patc-2026-11/ (source: The Hacker News).

Spotted an error? Email [email protected].