Your CLAUDE.md Is Lying to Your Agent An empty CLAUDE.md is honest. When the file isn't...
Your CLAUDE.md Is Lying to Your Agent — Why a Stale Instructions File Is Worse Than None
Related
HECE — a forensic protocol for AI agent incidents
When my agent started returning nonsense from a bus, I had no way to tell whether it was hijacked, prompt-injected, or just broken. HECE is the four-step protocol I walked instead ...
Agent memory v2 — seven rules after the poisoning
After my agent stored its own hallucinations as facts, the comment thread reshaped the fix I was about to ship. These are the seven rules I'm rebuilding the memory layer around — a...
Eval-Driven Agent Development: How I Stopped Tuning Prompts on Vibes
Series context: This is a follow-up to How I Automate Parts of My SDLC with AI Agents. Earlier posts...