Shipped Known Risk: What GPT-5.6 Sol's File Deletions Reveal About Agentic AI Governance
OpenAI's GPT-5.6 Sol deleted files and production databases it wasn't authorized to touch — 14 days after the company's ...
Read MoreAI Security
50 articles
PolinRider: North Korea's Open Source Supply Chain Campaign and What It Means for Healthcare DevSecOps
A North Korean supply chain campaign has compromised 108 open source packages and 1,951 GitHub repositories since Decemb...
Read MoreClaude Code Desktop: A Security Evaluation for Healthcare DevSecOps Programs
Claude Code Desktop is the most capable AI-assisted development environment available in a desktop app today — and it sh...
Read MoreAnthropic's Cyber Verification Program in Practice: From Approval to Real Defensive Work
Anthropic's CVP approved in one business day. Here's what the classifier actually unlocks, how Sonnet 5 and Opus 4.8 com...
Read MoreWhat the Scatter Spider Arrest Reveals About Windows Telemetry and PHI Risk on Clinical Workstations
A Scatter Spider member was caught partly through Windows Telemetry data. The same default-enabled collection mechanisms...
Read MoreDigital Nuclear Weapons: What the CIA Director's AI Analogy Gets Right, Gets Wrong, and What It Means for Healthcare Security
CIA Director Ratcliffe called frontier AI "akin to digital nuclear weapons" at the AWS Summit. Security analysts are spl...
Read MoreClaude Code's Hidden Fingerprint: A Security Analysis of the Steganographic Detection Mechanism
Anthropic's Claude Code was caught using steganographic Unicode encoding to silently fingerprint users routing through C...
Read MoreFable 5 Came Back Different: What the New Safeguard Taxonomy Means for Security Practitioners
Anthropic published the most detailed AI cybersecurity classifier taxonomy any frontier lab has produced — and it shows ...
Read MoreFable 5 Restored and the Jailbreak Severity Framework: Closing the Series, Opening the Governance Conversation
The 19-day Fable 5 suspension is over. But the more significant development is what emerged from it: a four-criteria jai...
Read More