@bman654/clodex 2.7.0 → 2.8.1

package/README.md

@@ -30,7 +30,7 @@ clodex claude                  # 5. launch Claude Code on an OpenAI model
 ```
 
 1. **Install** — puts the `clodex` command on your PATH.
-2. **Sign in** — opens a device-code OAuth flow for your ChatGPT/Codex plan; the token is stored in your OS credential store. (API-key users: `clodex providers add` instead.)
+2. **Sign in** — opens a device-code OAuth flow for your ChatGPT/Codex plan; the token is stored in your OS credential store. If your workspace admin has disabled device code authorization, add `--browser` to sign in through your browser instead. (API-key users: `clodex providers add` instead.)
 3. **Pick models** — an interactive manager for favorites (max 20) and short aliases like `sol` so you do not need to type the long names. Favorites drive the `/model` switch menu, proxy-mode routing, and the patcher.
 4. **Patch** *(optional but recommended for proxy mode)* — bakes your favorites and aliases into the Claude Code binary so they pass model validation, appear in `/model`, and report their real context windows. Re-run after each `claude` update; `clodex patch --restore` undoes it. This step is required if you want to use clodex-routed models as subagents via the Agent tool.
 5. **Launch** — starts Claude Code bridged to the model you choose.
@@ -205,6 +205,10 @@ Enabling this feature executes that JavaScript with your full user permissions;
 
 The module must default-export an array. Each site has a unique lowercase `id` and an `apply` function that returns the complete source and emits the generated `marker` exactly once when it applies:
 
+The `source` your `apply` receives is every JavaScript module in the binary joined together, with a ``/*clodex:module-boundary`;{}"]*/`` line between them. This is true of every Claude Code version — a pre-2.1.242 binary carries the bundle plus five small helper modules; a 2.1.242-or-later one carries the bundle split across roughly 1,370. Match and replace as you always have; just do not add or remove one of those lines, or the whole patch is refused. (The punctuation in the marker is there so that a wildcard bounded by a character class cannot run across a module boundary.)
+
+If your transform works by **position** rather than by matching an anchor, mind where it lands: text prepended to `source` goes into the first module and text appended to it goes into the last, and neither is necessarily the module that runs at startup. Anchor your edit to text you match, the way the examples below do.
+
 ```js
 export default [
   {
@@ -273,7 +277,7 @@ Two things worth knowing about the numbers:
 | --- | --- |
 | *(none)* | Provider hub wizard |
 | `add` | Add OpenAI or OpenCode Go with an API key, or sign in with ChatGPT |
-| `auth openai` | Sign in with ChatGPT/Codex-plan OAuth (device code) |
+| `auth openai` | Sign in with ChatGPT/Codex-plan OAuth (device code; `--browser` for workspaces that disable device codes) |
 | `list` | Show configured providers |
 | `remove <id>` | Remove a provider by id |
 | `refresh-models [id]` | Update cached model lists |

package/dist/{chunk-WZXTVKFX.js → chunk-J7WXOD2K.js}

@@ -1708,6 +1708,7 @@ export {
   getInstalledClaudeVersion,
   tcpListenerUrlHost,
   waitForTcpListenerCandidate,
+  waitForTcpListener,
   listenTcpServer,
   isDiscoveryDisabled,
   registerServerRuntimeState,
@@ -1722,4 +1723,4 @@ export {
   wrapperRequiresServer,
   computeWrapperEnv
 };
-//# sourceMappingURL=chunk-WZXTVKFX.js.map
+//# sourceMappingURL=chunk-J7WXOD2K.js.map

package/dist/claude-wrapper.js

@@ -7,7 +7,7 @@ import {
   readLiveServerRuntimeStates,
   waitForTcpListenerCandidate,
   wrapperRequiresServer
-} from "./chunk-WZXTVKFX.js";
+} from "./chunk-J7WXOD2K.js";
 
 // src/claude-wrapper.ts
 import { spawn } from "child_process";