OPNsense WireGuard VPN Gateway Failover with IVPN
Set up tiered gateway failover for WireGuard VPN tunnels on OPNsense, so if your primary exit server goes down, traffic automatically fails over to the next.
AI & Security Blog
Come on in! Welcome to my little corner in the internet. This is where I get to spill my thoughts about cybersecurity, technology, and digital privacy. Hopefully I can help you stay secure in the digital world.
Latest Posts
Why Context Windows Kill AI Agent Performance
Learn how context window degradation affects AI agents and why the Ralph Wiggum loop's genius lies in keeping agents in their 'smart zone' by resetting context between tasks
MCP Security: Why the Model Context Protocol is Insecure by Design
A technical deep-dive into MCP security vulnerabilities, attack vectors like tool poisoning and prompt injection, real-world incidents, and hardening strategies for securing your AI agent infrastructure.
Context Graphs: AI's Missing Decision Layer
Learn why context graphs are transforming enterprise AI by capturing decision traces · the missing layer between what happened and why it happened.
Smarter Scaffolding Beats Bigger Models
MIT's recursive language model technique offers unlimited context through clever infrastructure · no model changes needed. Why scaffolding innovations matter more than scaling.
Shell Metacharacters: How Command Injection Works
Learn how attackers exploit shell metacharacters for command injection. Covers prevention techniques, vulnerable code patterns, and secure alternatives for Python, Node.js, and more.
My Blog Got Hijacked (Sort Of)
A war story about discovering my domain was serving a gambling site—and how the 3-hour investigation revealed it wasn't DNS hijacking at all.
curl | bash is BAD
Learn why piping scripts to bash is risky, how attackers exploit this pattern, and safer alternatives for script installation.
The Great Firewall's Wallbleed: How a Memory Bug Exposed China's Censorship Machine
A deep dive into DNS poisoning, the Great firewall of China, and the Wallbleed vulnerability that allowed researchers to exfiltrate 5 billion packets from Chinese censorship infrastructure.
Anthropic Announces Rate Limits: How to Maximize Your AI Coding Productivity
Strategies to optimize your agentic AI usage after Anthropic's new rate limits. Learn to extend coding sessions and get production-ready software without burning through quotas.
Getting the Most Out of Claude Code: A Developer's Guide to AI-Assisted Programming
Master Claude Code with proven techniques, best practices, and workflows to boost your programming productivity while maintaining code quality.
Why "I Have Nothing to Hide" Won't Protect You from Russian Hackers
Recent events involving Russian military intelligence targeting ordinary people through pirated software prove this thinking dangerously wrong.