1 article on this topic.
Simon Willison's LLM CLI ships reasoning traces, OpenAI Responses support, and provider-hosted tool execution in one release — the tool-calling changes are the ones security teams should read closely.