OpenAI Codex CLI

Users are experiencing critical stability issues with long-lived CLI sessions, including data loss on exit, memory leaks causing system freezes, orphaned child processes, and crashes when handling long chat histories. These issues undermine trust in the tool for important work, as users risk losing progress and context when sessions become unstable.

Users are experiencing problems with model responses in the CLI where outputs are either unexpected or garbled/incomprehensible, making them difficult to interpret. This degrades the CLI experience and undermines trust in the tool's reliability. Users need clear, accurate, and readable responses from model queries in the command-line interface.