OpenAI Codex CLI 0.146.1 hardens auto-review for cyber-capable models

OpenAI

Tools official 1 src. ~1 min

Codex CLI's stable rust-v0.146.1 release backports safer automatic-review defaults for cyber-capable models and clarifies permission-change explanations shown in the terminal interface.

Why it matters

A narrow but security-relevant patch tightening auto-review behavior specifically for models flagged as cyber-capable.

Importance: 1/5

Narrow patch release with a single security-hardening change.

codex coding-agent cli security

Sources

official OpenAI Codex CLI rust-v0.146.1