Fetched from the account-scoped ChatGPT endpoint (chatgpt.com/backend-api/wham/usage) and rendered in Settings:

Plan and credits (balance, or unlimited). Plans render under their user-facing names: plus → Plus, prolite → Pro 5x, promax / pro → Pro 20x, free → Free, and likewise for Team / Business / Enterprise / Education; unrecognized identifiers pass through unchanged so new tiers stay visible.
Primary and secondary rate-limit windows with used-percent progress bars and localized reset times.
A limit reached notice; auto-refreshes every 60 s while connected, plus a manual refresh button.

Profile patch ID: codex-provider. Set values in the profile's cordis.patch.yml overlay.

Option	Values	Default	Notes
`transport`	`sse` \| `websocket` \| `websocket-cached` \| `auto`	`sse`	See Reliability-first transport above.
`streamIdleTimeoutMs`	positive integer	`300000`	Max idle interval while reading one response stream.
`timeoutMs`	positive integer	—	Optional overall request timeout.
`websocketConnectTimeoutMs`	positive integer	—	Optional WebSocket connect timeout.
`retryPolicy`	retry policy object	built-in	Request retry behavior (same schema as other DSH providers).
`credentialRef`	credential reference	`OPENAI_CODEX_OAUTH`	Harness credential slot holding the serialized OAuth state.
`ipv6CallbackBridge`	boolean	`true`	Relay the OAuth loopback callback for IPv6-preferred hosts.
`proactiveRefresh`	boolean	`true`	Refresh access tokens ahead of expiry.
`proxyMode`	`auto` \| `environment` \| `off`	`auto`	Also editable in the Settings page; applies after restart.

Example overlay:

- id: codex-provider
  name: '@jcy2387/dsh-codex-provider'
  config:
    transport: sse
    streamIdleTimeoutMs: 300000
    proxyMode: auto

@jcy2387/dsh-codex-provider

安装