Threat models, guardrails and red teaming for LLMs and the systems built around them.
A practical look at why AI-written code needs a verification harness rather than trust in a capable chat window, covering context rules, local gates, clean-machine CI, human review and production proof.
AI assistant
Ask about AI security, CISO strategy, compliance or anything I've written. Answers are generated by Claude only from published pages on this site, with numbered links to the sections they use. If the site doesn't cover something, it says so. Answers can still be wrong, so check the sources.