Prompt Injection Is the SQL Injection of AI AgentsYou can't prompt your way out of prompt injection. Here's how I design AI agents for containment instead of hope. If you built web apps in the 2000s, you remember SQL injection. Someone typed '; DROP14 min read·michaelpetler.hashnode.dev4h ago
Allow, Ask, or Block: The Week Agent Security Moved to the ActionThe security industry just agreed on something I've been building around for a while. Guard what the agent does, not just what it reads. [Image: 2026-11-08-cover.jpg. Alt: A worn industrial panel with8 min read·michaelpetler.hashnode.dev1E4d ago
Why My AI Agents Aren't Allowed to Make Things UpZero-fabrication rules, source-or-silence, and labeled example data: how I keep AI agents honest inside a real business.9 min read·michaelpetler.hashnode.devOct 27
The AI Agent Harness Matters More Than the ModelRead-only-first agents, permission-denied by design, and glass-house transparency: how I build governed AI agents at INASNAP, without the fear. Most AI coverage argues about the model. Which one is sm15 min read·michaelpetler.hashnode.dev12MAOct 26
