codex-multi-auth 2.15.0 → 2.17.0

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "codex-multi-auth",
-	"version": "2.15.0",
+	"version": "2.17.0",
   "description": "Install and operate codex-multi-auth for the official @openai/codex CLI with multi-account OAuth rotation, switching, health checks, and recovery tools.",
   "interface": {
     "composerIcon": "./assets/codex-multi-auth-icon.svg"

package/README.md

@@ -39,6 +39,14 @@ Use it when you need a local Codex CLI multi-account workflow with visible accou
 
 ---
 
+## Optional automatic subscription priming
+
+Run `codex-multi-auth account auto-prime <index> on` to opt an account into
+periodic first-use checks by the running CLI/app router (every 15 minutes).
+It completes a tiny response only for an unused personal subscription and
+consumes subscription quota. The default is off; use `off` to disable it.
+Manual `check` still needs `--prime`. See the [command reference](docs/reference/commands.md).
+
 ## Why Developers Use It
 
 `codex-multi-auth` makes local Codex account state visible and recoverable. Instead of one opaque auth file, you get a named account pool, deterministic account switching, health-aware selection, JSON diagnostics for automation, and safe repair commands for stale or damaged local state. The architecture is designed for personal development workflows: credentials stay local, runtime rotation is loopback-only, and official Codex install paths keep owning the `codex` command.
@@ -58,6 +66,10 @@ Use it when you need a local Codex CLI multi-account workflow with visible accou
 
 The package does not publish a global `codex` binary. Keep `codex` owned by the official OpenAI install path. Use `codex-multi-auth-codex ...` or `mcodex ...` only when you intentionally want this package's forwarding wrapper.
 
+With runtime rotation enabled, interactive `mcodex resume` lists saved CLI and desktop sessions across providers in the current directory. Use `mcodex resume --all` to include other directories, or `-C <directory>` to change the project scope. Select a session to resume it through multi-auth; Escape cancels. Discovery uses the official local Codex app-server and does not rewrite session metadata. Desktop and CLI must use the same `CODEX_HOME` for shared history.
+
+Explicit session IDs, `--last`, remote connections, non-interactive invocations, and the in-session `/resume` command retain native Codex behavior. The cross-provider picker requires an official Codex version supporting the app-server `thread/list` API.
+
 ---
 
 <details open>
@@ -417,7 +429,7 @@ codex-multi-auth doctor --json
 
 ## Release Notes
 
-- Current stable: [docs/releases/v2.15.0.md](docs/releases/v2.15.0.md) — install via `npm i -g codex-multi-auth`
+- Current stable: [docs/releases/v2.17.0.md](docs/releases/v2.17.0.md) — install via `npm i -g codex-multi-auth`
 - Previous stable: [docs/releases/v2.8.3.md](docs/releases/v2.8.3.md)
 - Previous stable: [docs/releases/v2.8.2.md](docs/releases/v2.8.2.md)
 - Previous stable: [docs/releases/v2.7.1.md](docs/releases/v2.7.1.md)
@@ -461,3 +473,12 @@ MIT License. See [LICENSE](LICENSE).
 - You assume responsibility for your own usage and compliance.
 
 </details>
+
+### Optional native desktop binding
+
+Native desktop features such as Remote Control pairing need the real desktop
+account identity. Use `codex-multi-auth rotation bind-app --native` to preserve
+that login while routing inference through managed accounts. Add
+`--catalog-account <index>` to use one account's live model catalog. This mode
+requires file-backed desktop credentials; see the [rotation command reference](docs/reference/commands.md#native-provider-binding-opt-in)
+for setup, limitations, and rollback.

package/config/README.md

@@ -17,8 +17,8 @@ These files are optional model/provider templates for host/plugin configurations
 
 ## Defaults Included
 
-- Current documented OpenAI/Codex model families: GPT-6 Astra (`gpt-6-astra`, `gpt-6-astra-aeon`), then `gpt-5.6-sol`/`terra`/`luna`, `gpt-5.5`, `gpt-5.4`, and `gpt-5.3-codex`
-- The Astra entries carry `limit.context: 272000`, OpenAI's stable Codex context value at launch. The GPT-5.6 entries still carry the pre-cut `372000` and were left alone here; change them only alongside a deliberate review, since the value feeds a picker rather than the context budget guard (which refuses to estimate either family)
+- Current documented OpenAI/Codex model families: GPT-6 Astra (`gpt-6-astra`, `gpt-6-astra-aeon`), GPT-6 Sol (`gpt-6-sol`) and GPT-6 Luna (`gpt-6-luna`), then `gpt-5.6-sol`/`terra`/`luna`, `gpt-5.5` and `gpt-5.5-pro`. Retired models (`gpt-5.4*`, `gpt-5.3-codex` and older) were removed; their ids still resolve to a replacement, see `docs/configuration.md`
+- The GPT-6 entries carry `limit.context: 272000`, the `context_window` in the upstream Codex catalog. The GPT-5.6 entries still carry the pre-cut `372000` and were left alone here; change them only alongside a deliberate review, since the value feeds a picker rather than the context budget guard (which refuses to estimate either family)
 - The Daybreak cyber models resolve at the code level but are deliberately absent from these templates, matching their hidden visibility in the upstream Codex picker
 - `store: false`
 - `include: ["reasoning.encrypted_content"]`