Bug in Google's Gemini AI Panel Opens Door to Hijacking

2026-03-04T20:36:38Z1de303ba5f30c32415595c7e2171b71bc677279ce999f61a5fda2bee7baca67a
CVE-2026-20127CVE-2026-2329ai-agentsai-securitycisco-sd-wanclaude-codefortigategeminigooglegrandstreamlazarus-groupmalicious-reposmedusamfa-bypassnextjsnpmopenclawphishingprivacyprivilege-escalationransomwarereact2shellstarkillersupply-chainthreat-intelligence

What happened

Collection of DarkReading headlines (Feb–Mar 2026) covering multiple high-impact security incidents and trends: a vulnerability in Google’s Gemini AI panel that could allow privilege escalation and privacy violations; a maximum-severity Cisco SD‑WAN zero‑day (CVE-2026-20127) reportedly exploited for three years; a critical Grandstream VoIP bug (CVE-2026-2329) granting unauthenticated root access; mass compromises of FortiGate devices using generative AI; widespread ransomware activity (including Medusa linked to Lazarus) hitting healthcare and other sectors; supply‑chain attacks (malicious npm

Why it matters

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

Evidence and limitations

Source ID
darkreading
Record identifier
1de303ba5f30c32415595c7e2171b71bc677279ce999f61a5fda2bee7baca67a
Enrichment time
2026-03-04T20:36:38Z
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.

Record · Bug in Google's Gemini AI Panel Opens Door to Hijacking · Baitaphish