Articles
No articles match. Clear filters
article
The Math Problem AI Just Changed for Security Testing
Security testing has always been an economics problem: defenders can only test what they can afford to test. AI changed the cost of testing, and that changes which assumptions about coverage still hold.
22 Mar, 2026
Open document : The Math Problem AI Just Changed for Security Testingarticle
Four Generations of Broken Promises: Why AI SOC Agents Might Actually Be Different
Three generations of security tooling promised to fix the analyst shortage. All three failed. The fourth - AI SOC Agents - operates on a different principle, but CISOs who ignore twenty years of lessons will repeat the same expensive mistakes.
18 Mar, 2026
Open document : Four Generations of Broken Promises: Why AI SOC Agents Might Actually Be Differentarticle
I Scanned 152 Files of My Own AI-Generated Code for Invisible Unicode Malware
GlassWorm hid malicious code in invisible Unicode characters. I scanned 152 files of my own AI-generated code to see whether the same trick was already sitting in my repositories - and built a pre-commit check from what I found.
17 Mar, 2026
Open document : I Scanned 152 Files of My Own AI-Generated Code for Invisible Unicode Malwarearticle
Claude Code Has Two New CVEs - Here's What They Exploit and How to Harden Your Setup
Check Point disclosed two Claude Code CVEs exploiting hooks and MCP config files. Here's what the attack chains look like and how to harden your environment.
04 Mar, 2026
Open document : Claude Code Has Two New CVEs - Here's What They Exploit and How to Harden Your Setuparticle
How I Made Claude Code Safer (And You Can Too)
Claude Code validates which tools can run, not what they write. That gap cost me a crashed project and a malformed config file - so I built a plugin that validates content before it hits disk, and it turned out to teach Claude to stop repeating the same mistakes.
10 Feb, 2026
Open document : How I Made Claude Code Safer (And You Can Too)experiment
I Ran 849 Tests on AI Context Files. Here's What Actually Works.
849 controlled tests across three corpus sizes and five folder structures found that a single flat folder with descriptive filenames beats nested hierarchies at every scale - and that keyword or summary indexes actively hurt accuracy once a corpus passes 300K words.
05 Feb, 2026
Open document : I Ran 849 Tests on AI Context Files. Here's What Actually Works.