New ChatGPT Lockdown Mode Limits Tools That Could Enable Data Exfiltration

2026-06-08T01:24:08Z20d149ac715e35ad52497d724988b77423057aa9d3577e30186f7dd27d0dd56c
active-exploitationai-fuzzingandroidchromiumcloud-compromisedosgithubmacosmalwarenpmpatchingphishingproxy-networkrcespywaresupply-chainthreat-actorweb-shellwormzero-day

What happened

A batch of high-impact incidents and vulnerabilities affecting diverse platforms was reported: OpenAI rolled out a ChatGPT Lockdown Mode to mitigate prompt-injection data exfiltration, while researchers disclosed Bright Data SDK abuse that turns consumer devices (including smart TVs) into residential proxy exit nodes. Multiple actively exploited and high-severity flaws were highlighted — notably SolarWinds Serv‑U DoS (CVE-2026-28318), Cisco Catalyst SD‑WAN Manager (CVE-2026-20245), Cisco Unified CM (CVE-2026-20230), and a critical Everest Forms Pro RCE (CVE-2026-3300) — alongside a Redis RCE (

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
the_hacker_news
Record identifier
20d149ac715e35ad52497d724988b77423057aa9d3577e30186f7dd27d0dd56c
Enrichment time
2026-06-08T01:24:08Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.