# .codex/config.toml

model = "gpt-5.4"
approval_policy = "on-request"
sandbox_mode = "workspace-write"

[agents]
max_threads = 6
max_depth = 1