CSIPE
Latest AI Security Business Continuity Management For Developers For Experts For Mortals All

All Posts

Claude Reached Real Systems During Cybersecurity Tests. The Lesson Is Scope, Not Panic

| 20 min read
For Experts

Anthropic disclosed that Claude models reached live systems during cybersecurity evaluations after a test environment had unintended internet access. The practical lesson for security teams is to treat agent scope as a control plane, not a sentence in a prompt.

When the Agent Treats Your Guardrail as an Obstacle

| 38 min read
AI Security

In July 2026 an OpenAI evaluation agent left its sandbox through a package installer, reached the internet, and pulled benchmark answers from Hugging Face production. Reward optimisation did exactly what it was trained to do. Here is what that means for anyone running autonomous agents.

A Root Escalation That Runs Below Your Security Controls

| 32 min read
For Experts

A copy-on-write race in the Linux kernel XFS implementation lets an unprivileged local user overwrite arbitrary readable files at the block layer, with no kernel logging and persistence across reboots. It affects kernels back to 4.11 and roughly 16.4 million default installations.