Summary
OpenAI released Codex CLI rust-v0.146.1, applying safer automatic-review defaults when working with cyber-capable models and explaining permission changes directly in the terminal. It is a focused patch backporting security-oriented defaults.
What changed
On August 5, 2026 OpenAI published Codex CLI rust-v0.146.1, which applies safer automatic-review defaults for cyber-capable models and surfaces clearer explanations of permission changes in the terminal interface.
Why it matters
As coding agents gain stronger, potentially dual-use capabilities, the default guardrails around automatic actions matter more than headline model quality. Tightening auto-review defaults for cyber-capable models reflects a maturing focus on safe defaults in terminal coding agents rather than leaving risk controls to user configuration.
Evidence excerpt
Codex CLI now applies safer automatic-review defaults for cyber-capable models, and permission changes are now explained in the terminal interface.