OpenAI Confirms AI Agent Wiki Incident and Disclosure Gap
OpenAI acknowledged an AI agent wiki incident involving roughly 18,000 posts and said it is developing broader disclosure rules for unexpected behavior.
Read more →Cybersecurity
Latest articles in Cybersecurity
OpenAI Pauses Astra Work Over Critical Cyber Capability
OpenAI paused some Astra work after early tests could not rule out critical cyber capability, prompting stricter isolation, monitoring and external review.
Read more →
China-Linked Hackers Exploit VMware ESXi Zero-Days to Escape Virtual Machines
Chinese-linked threat actors exploit three VMware ESXi zero-days (CVE-2025-22224, CVSS 9.3) to escape VMs and compromise hypervisors using sophisticated MAESTRO toolkit.
Read more →
AI-Generated Phishing Emails Are 4.5 Times More Effective, Microsoft Warns
Microsoft reveals AI-generated phishing emails achieve 54% click rates vs 12% for traditional phishing - a 4.5x increase making campaigns 50x more profitable for attackers.
Read more →
Yahoo and AOL Hit by Major Outage as Edge Servers Block Thousands of Users
Yahoo Mail and AOL services down for 27,700+ users worldwide with 'Edge: Too many requests' error. Rate limiting misconfiguration suspected as cause.
Read more →
X Suffers Major Global Outage as Cloudflare Errors Block 96,000 Users
X platform experiences major global outage with 96,000+ users blocked by Cloudflare errors, marking second disruption this week amid recurring infrastructure issues.
Read more →
OWASP AI Security Guide: 300 Pages of Practical Protection Guidance
OWASP releases comprehensive 300-page AI Security Guide with practical guidance covering complete AI lifecycle security from training through operations.
Read more →
Microsoft Releases Framework for Building Secure AI Agents at Enterprise Scale
Microsoft Cloud Adoption Framework provides enterprise guidance for building secure AI agents with Azure Foundry controls and standardized governance processes.
Read more →
Why Your AI Agent Needs Different Monitoring: Traditional Tools Fall Short
Michael Hanne argues traditional monitoring fails AI agents. Non-deterministic behavior and behavioral drift attacks require purpose-built observability approaches.
Read more →
The Map Is Not the Territory: Understanding Agent Tool Trust Boundaries
Niyikiza research explores AI agent tool trust boundary gaps where tool descriptions diverge from actual capabilities, creating exploitable security vulnerabilities.
Read more →
Page 1 of 2
Next »