I meant to do that! AI vendors shrug off responsibility for vulns

2026-04-19T19:23:57Z399d377edbcfc475a4c61d3cfbe9272af3c246f6d0151f3f50511ce5c6d11628
AI securityActiveMQAnthropicApacheCISAExcelFortinetGitHub agentsKEVLLM vulnerabilityMicrosoft Patch Tuesdaycritical-vulnerabilitiesdata breachmacOS social engineeringprompt injectionransomwarezero-day

What happened

A roundup of security news highlighting multiple active and high-impact issues: US CISA ordered rapid patching for a newly-exploited 13‑year‑old Apache ActiveMQ flaw and a 17‑year‑old critical Excel bug; Fortinet sandbox vulnerabilities rated critical could allow unauthenticated bypass and code execution; Microsoft released a large Patch Tuesday fixing numerous flaws (including exploited ones). The feed also covers AI/LLM security failures and design issues (Anthropic MCP, Mythos/Project Glasswing, Claude-built exploits), prompt‑injection and agent hijacking against GitHub‑integrated agents, a

Why it matters

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

Evidence and limitations

Source ID
theregister_security
Record identifier
399d377edbcfc475a4c61d3cfbe9272af3c246f6d0151f3f50511ce5c6d11628
Enrichment time
2026-04-19T19:23:57Z
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.