Configurable TUI status bar fields for model, effort, context, and rate-limit usage

Summary

Please add a supported way to customize the Codex terminal UI status bar so users can show model/runtime usage fields directly in the existing bottom bar.

Example desired status line:

gpt-5.4 | effort:medium | ctx:18% | 5h:56%

Current behavior

The Codex TUI currently shows useful basics such as the active model and current folder, but I do not see a user-facing config surface for formatting the status bar.

Locally verified on:

codex-cli 0.142.5

The installed package is a native binary launched through @openai/codex, so durable local customization is not practical. Local scripts can read session JSONL events, but they cannot render inside the built-in Codex status bar.

Proposed config

One possible shape:

[tui.status_bar]
format = "{model} | effort:{effort} | ctx:{context_percent}% | 5h:{rate_limit_5h_percent}%"

A minimal implementation could support a fixed set of placeholders:

{model}
{effort}
{cwd}
{context_percent}
{context_used_tokens}
{context_window_tokens}
{rate_limit_5h_percent}
{rate_limit_5h_resets_at}

Why this belongs in Codex

Codex already records the underlying data in session/runtime events:

• turn_context.payload.model
• turn_context.payload.effort
• token_count.info.last_token_usage / total_token_usage
• token_count.info.model_context_window
• token_count.rate_limits.primary.used_percent for the 300-minute window

Because the TUI owns the status bar, this is cleaner and more durable as a first-class Codex feature than as a local wrapper, prompt instruction, polling script, or patch to the installed package.

User value

This gives users immediate awareness of model, reasoning effort, context pressure, and 5-hour usage pressure without spending tokens, adding assistant output noise, or running a separate sidecar UI.