AI reliability

A catalog of design patterns for building agentic systems that verify outputs with evidence instead of relying on self-review. The page explains two main failure modes and presents structured patterns for context, verification, and orchestration.