Contributing to Open Source Again, With AI
How I used Claude Code to build an MCP server for Tiki Wiki and submit my first open source contribution in years.
Notes on IAM, cybersecurity, AI, and emerging technologies
Recent log
Scaling OpenClaw's Agentic AI to Multi-User Environments with OpenFGA Authorization
Adding fine-grained authorization to OpenClaw's agentic AI platform using OpenFGA for multi-user environments.
My Son's Spring Break Linux Experiment (and the WiFi Rabbit Hole That Almost Killed It)
My son wanted to try Linux over spring break for better FPS. What followed was a WiFi driver rabbit hole that nearly ended the experiment.
Recent research
The End of NTLM and the Kerberos Reckoning
Microsoft is dismantling NTLM through a three-phase deprecation plan while Kerberos faces its own expanding attack surface. A comprehensive look at the threat landscape, vendor responses, and the modern alternatives enterprises must adopt.
The Audit Blind Spot: Why Workload Identities Escaped Review and Why That Era Is Ending
Most organizations cannot pass a rigorous audit of their non-human identity controls today. PCI DSS 4.0.1, DORA, and NYDFS now explicitly require service account access reviews, yet the average enterprise still has 60% of its AWS IAM access keys older than one year.