Anthropic Finds 22 Firefox Vulnerabilities Using Claude Opus 4.6 AI Model
2026-03-07T13:24:13Z•d2abd22a33e25361982389b7c670dfa1557de29ce0927f5ae963d42f6e8e90b2
AI-assisted-vulnerability-discoveryAPTAsyncRATCISA-KEVCiscoClickFixCoruna-iOS-exploit-kitCyberStrikeAIDindoorFortiGateHikvisionLumma-StealerMuddyWaterRATVMwareVOID#GEISTXWormXenoRATactive-exploitationai-generated-malwareanthropicfirefoxpackagist-malwaresupply-chaintransparent-tribe
What happened
A batch of mid-March 2026 cybersecurity reports highlights rapid AI adoption across both defensive and offensive workflows, multiple active exploitation events, and diverse malware/APT activity. Key items: Anthropic used its Claude Opus 4.6 model to discover 22 Firefox vulnerabilities addressed in Firefox 148; transparent use of AI by threat actor Transparent Tribe to mass-produce implants in lesser-known languages; multi-stage VOID#GEIST campaigns delivering XWorm, AsyncRAT and Xeno RAT; MuddyWater’s new Dindoor backdoor; active exploitation and CISA KEV additions for high-severity flaws (not
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- the_hacker_news
- Record identifier
- d2abd22a33e25361982389b7c670dfa1557de29ce0927f5ae963d42f6e8e90b2
- Enrichment time
- 2026-03-07T13:24:13Z
- 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.