Safety · Hacker News ·
Codex starts encrypting sub-agent prompts
An issue in OpenAI’s Codex repository reports that the tool now encrypts prompts sent to sub-agents. The change concerns the handling of prompt data and has prompted discussion among users.