dsh-codex-subscription 1.14.2 → 1.14.4

package/README.en.md

@@ -22,16 +22,6 @@ No OpenAI API key or Codex CLI. Models, search, quota, and image generation stay
   <img src="https://raw.githubusercontent.com/WSL043/dsh-codex-subscription/main/docs/assets/readme-hero-en.webp" width="900" alt="Your Codex subscription inside DSH: models, web search, quota and safe reset, image generation, and Fast mode">
 </p>
 
-## What's included in 1.14.2
-
-- GPT-6 Astra now supports model-catalog context settings: Standard preserves the catalog default, Extended uses 872K, and Custom is bounded to 128K–872K.
-
-### From 1.14.1
-
-- The composer shows both 5-hour and weekly quota windows when returned by the service, so neither window is hidden.
-- Image edits validate attachment references, normalize bare SHA-256 digests, and resolve the actual attachment metadata from the current session. Paths produce an actionable error asking the model to obtain the correct reference with `read_image` and retry.
-- Location edits attach the clean source, a numbered annotation image, and coordinate instructions so the model can match each requested change to its location.
-
 ## Three-step start
 
 1. **Install the plugin.** Run the standard DSH bundle command:
@@ -50,14 +40,14 @@ DSH-Portable exposes the same standard plugin command, so the command above also
 | Capability | What you get |
 | --- | --- |
 | **Subscription models** | Sign in to ChatGPT and use Codex without an OpenAI API key or Codex CLI |
-| **Recoverable and diagnosable** | Sign-in state reconciles automatically; Settings can create a support report without credentials or account identifiers |
+| **Recoverable and diagnosable** | Sign-in state reconciles automatically; failed reads can be retried in place, while timeouts and stale account responses cannot overwrite current state; Settings can create a support report without credentials or account identifiers |
 | **Visible quota** | Keep backend-provided standard Codex, Spark, and other limits separate |
 | **Composer quota** | Choose a compact percentage, progress bar, or no inline quota display |
 | **Safe quota reset** | See each reset credit separately and deliberately try one with a cooldown and acknowledgement |
 | **Subscription search** | Explicitly route search globally through DSH default search or the signed-in Codex subscription |
 | **Codex image generation and editing (Beta)** | Generate without references, or explicitly edit one selected image; preview, zoom, annotate regions, download the original, and get the original host path for new or edited images |
 | **Fast mode** | Switch between Standard and Fast directly in the composer |
-| **Model-aware context** | Keep catalog defaults, use each model's supported extended window, or enter a full numeric token limit for each model |
+| **Model-aware context** | Keep catalog defaults, use each model's supported extended window, or enter a full numeric token limit for each model; Settings refreshes the model directory on open, account changes, and connection resets without overwriting an unsaved draft |
 | **Headless runs** | Use the same signed-in Codex provider for one-shot DSH tasks that print their answer and exit |
 
 These capabilities reuse the same local ChatGPT sign-in. Subscription routing failures stay visible and never silently switch to another paid route.
@@ -181,7 +171,7 @@ A new image request does not silently include earlier images. GPT Image 2 can ta
   <img src="https://raw.githubusercontent.com/WSL043/dsh-codex-subscription/main/docs/assets/image-preview-annotations-en.png" width="800" alt="Generated image, region note, and continue editing inside the DSH Image Viewer">
 </p>
 
-The screenshot above illustrates image viewing and on-image notes. Version 1.14.0 includes the same basic experience; available buttons can vary with the image and installed viewer version.
+The screenshot above illustrates image viewing and on-image notes; available buttons can vary with the image and installed viewer version.
 
 ### Composer speed
 

package/README.md

@@ -23,16 +23,6 @@
   <img src="https://raw.githubusercontent.com/WSL043/dsh-codex-subscription/main/docs/assets/readme-hero.webp" width="900" alt="Codex 订阅直接用在 DSH：订阅模型、联网搜索、额度与安全重置、图片生成和高速模式">
 </p>
 
-## 1.14.2 重点变化
-
-- GPT-6 Astra 支持模型目录上下文设置：标准模式保留目录默认值，扩展模式使用 872K，自定义模式限制为 128K–872K。
-
-### 1.14.1 更新回顾
-
-- 输入框同时显示服务端返回的 5 小时与每周额度，避免其中一个额度窗口被隐藏。
-- 图片编辑会检查附件引用，自动补全裸 SHA-256，并从当前会话取回真实附件元数据；误填路径时提示模型通过 `read_image` 获取正确引用并重试。
-- 部位修改会同时附上干净原图、编号标注图和坐标说明，让模型能够对应每条修改要求。
-
 ## 三步开始
 
 1. **安装插件**：直接运行 DSH 标准 bundle 命令：
@@ -51,14 +41,14 @@ DSH-Portable 也提供相同的标准插件命令，因此同样使用上面的
 | 能力 | 用户得到什么 |
 | --- | --- |
 | **订阅模型直连** | 登录 ChatGPT 后直接使用 Codex，不需要 OpenAI API Key 或 Codex CLI |
-| **可恢复、可诊断** | 登录状态会自动对账；设置页可生成不含凭据和账号标识的支持报告 |
+| **可恢复、可诊断** | 登录状态会自动对账；读取失败时可在原处重试，超时和旧账号响应不会覆盖当前状态；设置页可生成不含凭据和账号标识的支持报告 |
 | **额度可见** | 普通 Codex、Spark 等服务端实际返回的额度分开显示 |
 | **输入框额度** | 可选择紧凑百分比、进度条或完全关闭输入框额度显示 |
 | **安全额度重置** | 每张重置卡单独显示，并通过冷静期和知情确认主动尝试重置 |
 | **订阅搜索** | 可将全部模型的搜索明确路由到 DSH 默认搜索或已登录的 Codex 订阅 |
 | **Codex 图片生成与编辑（Beta）** | 可无参考图全新生成，也可明确选择会话图片继续编辑；支持预览、缩放、区域备注、下载原图，并为新生成或编辑的图片提供 DSH 主机上的原图路径 |
 | **高速模式** | 直接在输入框切换标准或高速，无需离开当前会话 |
-| **模型感知上下文** | 可保留目录默认值、按模型启用扩展窗口，或为每个模型填写完整数字 Token 上限 |
+| **模型感知上下文** | 可保留目录默认值、按模型启用扩展窗口，或为每个模型填写完整数字 Token 上限；设置页打开、账号切换和连接重置后会刷新模型目录，失败时可重试且不会覆盖未保存的草稿 |
 | **Headless 任务** | 使用同一份已登录的 Codex Provider 运行一次性 DSH 任务，输出答案后自动退出 |
 
 这些能力共用同一份本机 ChatGPT 登录。订阅路由失败时会明确报错，不会静默切换到其他付费路由。
@@ -179,7 +169,7 @@ ChatGPT 返回可用重置卡时，设置页会把每张卡分别显示为紧凑
   <img src="https://raw.githubusercontent.com/WSL043/dsh-codex-subscription/main/docs/assets/image-preview-annotations.png" width="800" alt="DSH 图片查看器中的生成图、区域备注和继续编辑">
 </p>
 
-上图展示图片查看与图上备注的基本交互。1.14.0 已内置相同的基础体验；具体按钮会随图片和所安装的查看器版本变化。
+上图展示图片查看与图上备注的基本交互；具体按钮会随图片和所安装的查看器版本变化。
 
 ### 输入框速度