Support cost and budget metrics in TUI status_line

[thinker0]

Codex TUI currently supports token/context/rate-limit status line items such as context-used, used-tokens, total-input-tokens, total-output-tokens, five-hour-limit, and weekly-limit.

Please add native cost and budget status line items so users can monitor estimated spending directly in the bottom status bar.

Desired status line items:

• estimated-cost-current-session
• estimated-cost-today
• estimated-cost-month
• billing-budget-limit
• billing-budget-used
• billing-budget-remaining
• billing-budget-percent
• rate-limit-reset-time
• five-hour-limit-percent
• weekly-limit-percent

Example display:

GPT-5.5 | team-brain (main) | $34.7 | today:$34.7 mo:$5108.1 | ctx:23.0% | 5h:3.0% reset:4h43m | 7d:21.0%

Requirements:

• Cost should be calculated using the active model provider's actual pricing when available.
• If exact billing data is unavailable, Codex should clearly mark values as estimated.
• Cached input, cache write, output, and long-context multipliers should be handled separately.
• For custom providers, allow users to configure per-model rates in config.toml.
• Status line should expose both compact and detailed forms.
• Values should not leak API keys, organization IDs, or account-sensitive billing metadata.

Why:

Users already see token and context usage in the TUI, but cannot quickly understand cost impact. For long-context coding sessions, cost awareness is important for budget control and model selection.