HacktronAI reported chaining two vulnerabilities on July 25, 2026, to compromise OpenAI employees’ ChatGPT and Codex accounts and potentially reach connected services such as GitHub, Slack, and email....