Declan Ferreira

AI agents need permission controls designed for their speed and scope, not human role models.

Fast detection and recovery paths for OpenClaw's most destructive production failure modes.

LLM agents break idempotency by regenerating requests instead of replaying them.

Silent failures cost more than crashes—here's how to catch them.

Autonomous sales agents face prompt injection attacks through untrusted external content.