@mofeng2223/dsh-claude-provider 0.2.3 → 0.3.1

package/README.md

@@ -2,7 +2,11 @@
 
 English | [简体中文](./README.zh-CN.md)
 
-DeepSeek Harness's generic reasoning controls do not fully match the request parameters expected by newer Claude models. A selected level can therefore fail with HTTP 400, or be silently mapped to another effective level, such as Max behaving as High. `@mofeng2223/dsh-claude-provider` adds an explicit Claude provider type and sends the reasoning parameters that each configured Claude model expects.
+~~DeepSeek Harness's general-purpose reasoning controls did not fully match the request parameters required by newer Claude models. As a result, selecting a reasoning level could either produce an HTTP 400 error or be silently mapped by the adapter to a different effective level—for example, Max in the interface could actually be sent as High. `@mofeng2223/dsh-claude-provider` added an explicit Claude provider type and sent the correct reasoning parameters for each Claude model.~~
+
+DeepSeek Harness 0.1.0-rc.8 now supports reasoning-effort parameters for Claude models, but they still need to be configured manually in `settings.yaml`. Therefore, this plugin no longer intercepts or rewrites model requests at runtime. Instead, it writes the corresponding native settings to `settings.yaml` when a Claude provider is saved through the front end.
+
+The plugin retains its dedicated Claude provider configuration page, so users do not need to edit `settings.yaml` manually. Anthropic-native model discovery, automatic defaults for common Claude models, and other features not yet available in DeepSeek Harness are also retained.
 
 ## What this plugin does
 
@@ -28,7 +32,11 @@ DeepSeek Harness's generic reasoning controls do not fully match the request par
    - Four levels: Low, Medium, High, and Max
    - Toggle: On or Off
 
-   New models default to five levels. Claude providers default to High, while toggle-only models default to On. For adaptive-thinking models, the plugin converts the selection to Claude's `thinking.type: adaptive` and `output_config.effort` request format instead of allowing the adapter to collapse or reject the selected level.
+   New models default to five levels, and Claude providers default to High.
+
+   ~~For adaptive thinking models, the plugin converted the selected level into Claude's `thinking.type: adaptive` and `output_config.effort` request format, preventing adapter rejections or silent downgrades.~~
+
+   For adaptive thinking models, the plugin writes the selected level to DSH's native configuration when saved.
 
 <p align="center">
   <img src="./docs/images/model-defaults.en.jpg" alt="Claude model capacity defaults and reasoning modes" width="580">
@@ -44,7 +52,7 @@ DeepSeek Harness's generic reasoning controls do not fully match the request par
 
 5. **Leaves every other provider unchanged**
 
-   All discovery, defaults, reasoning controls, and request rewriting are limited to Provider IDs explicitly created as **Claude Providers**. DeepSeek Harness's built-in providers and ordinary custom providers keep their original behavior, even when they use `anthropic-messages` or expose a `claude-*` model ID.
+   Discovery, defaults, and reasoning controls are limited to Provider IDs explicitly created as **Claude Providers**. DeepSeek Harness's built-in providers and ordinary custom providers keep their original behavior, even when they use `anthropic-messages` or expose a `claude-*` model ID.
 
 ## Install
 
@@ -98,7 +106,7 @@ npx @deepseek-ai/dsh plugin --profile web remove @mofeng2223/dsh-claude-provider
 npx @deepseek-ai/dsh plugin --profile headless remove @mofeng2223/dsh-claude-provider
 ```
 
-Uninstalling the plugin does not delete `~/.dsh/settings.yaml` or stored credentials. Existing Claude providers remain configured as ordinary custom `anthropic-messages` routes, but lose the Claude-specific interface, model discovery, defaults, reasoning controls, and adaptive request conversion supplied by this plugin.
+Uninstalling the plugin does not delete `~/.dsh/settings.yaml` or stored credentials. Existing Claude providers remain ordinary custom `anthropic-messages` routes. Their `reasoningEfforts`, default `reasoning`, and `compat.forceAdaptiveThinking` fields are native RC8 settings, so adaptive thinking and the front-end reasoning-effort selector continue to work. Only the dedicated Claude add/edit UI, model discovery, and recorded defaults disappear.
 
 ## License
 

package/README.zh-CN.md

@@ -2,7 +2,11 @@
 
 [English](./README.md) | 简体中文
 
-DeepSeek Harness 的通用推理强度控制与新版 Claude 模型要求的请求参数并不完全匹配，因此选择某个档位后，可能直接出现 HTTP 400，也可能被适配器静默映射成另一个实际档位，例如界面选择 Max，实际请求却只相当于 High。`@mofeng2223/dsh-claude-provider` 增加了一个明确的 Claude 提供方类型，并按照每个 Claude 模型的配置发送正确的推理参数。
+~~DeepSeek Harness 的通用推理强度控制与新版 Claude 模型要求的请求参数并不完全匹配，因此选择某个档位后，可能直接出现 HTTP 400，也可能被适配器静默映射成另一个实际档位，例如界面选择 Max，实际请求却只相当于 High。`@mofeng2223/dsh-claude-provider` 增加了一个明确的 Claude 提供方类型，并按照每个 Claude 模型的配置发送正确的推理参数。~~
+
+DeepSeek Harness 0.1.0-rc.8 已支持 Claude 系列模型的推理强度参数，但仍需要在 `settings.yaml` 中手动配置。因此，本插件不再在运行时拦截或改写模型请求，而是在用户通过前端保存 Claude 提供方时，将相应的原生配置正确写入 `settings.yaml`。
+
+插件仍保留专用的 Claude 提供方配置页面，用户无需手动编辑 `settings.yaml`。Anthropic 原生模型探测、常用 Claude 模型参数自动填写等官方尚未实现的功能也继续保留。
 
 ## 插件功能
 
@@ -28,7 +32,11 @@ DeepSeek Harness 的通用推理强度控制与新版 Claude 模型要求的请
    - 四档：Low、Medium、High、Max
    - 开关：On、Off
 
-   新增模型默认使用五档。Claude 提供方默认选择 High，仅支持开关的模型默认选择 On。对于 adaptive thinking 模型，插件会把所选档位转换成 Claude 使用的 `thinking.type: adaptive` 与 `output_config.effort` 请求格式，避免适配器拒绝请求或把档位降级。
+   新增模型默认使用五档，Claude 提供方默认选择 High。
+
+   ~~对于 adaptive thinking 模型，插件会把所选档位转换成 Claude 使用的 `thinking.type: adaptive` 与 `output_config.effort` 请求格式，避免适配器拒绝请求或把档位降级。~~
+
+   对于 adaptive thinking 模型，插件会在保存时将所选档位写入 dsh 原生配置。
 
 <p align="center">
   <img src="./docs/images/model-defaults.jpg" alt="Claude 模型容量预填与思考模式设置" width="580">
@@ -44,7 +52,7 @@ DeepSeek Harness 的通用推理强度控制与新版 Claude 模型要求的请
 
 5. **不影响其他提供方**
 
-   模型探测、参数预填、推理档位和请求转换只对明确创建为“Claude 提供方”的 Provider ID 生效。DeepSeek Harness 的内置提供方和普通自定义提供方会保持原有行为，即使它们同样使用 `anthropic-messages` 协议，或者包含 `claude-*` 模型 ID，也不会被插件修改。
+   模型探测、参数预填和推理档位配置只对明确创建为“Claude 提供方”的 Provider ID 生效。DeepSeek Harness 的内置提供方和普通自定义提供方会保持原有行为，即使它们同样使用 `anthropic-messages` 协议，或者包含 `claude-*` 模型 ID，也不会被插件修改。
 
 ## 安装
 
@@ -98,7 +106,7 @@ npx @deepseek-ai/dsh plugin --profile web remove @mofeng2223/dsh-claude-provider
 npx @deepseek-ai/dsh plugin --profile headless remove @mofeng2223/dsh-claude-provider
 ```
 
-卸载插件不会删除 `~/.dsh/settings.yaml` 或已经保存的凭据。已有的 Claude 提供方会继续作为普通的 `anthropic-messages` 自定义路由保留，但会失去本插件提供的 Claude 专用界面、模型探测、参数预填、推理档位和 adaptive 请求转换。
+卸载插件不会删除 `~/.dsh/settings.yaml` 或已经保存的凭据。已有的 Claude 提供方会继续作为普通的 `anthropic-messages` 自定义路由保留；`reasoningEfforts`、默认 `reasoning` 和 `compat.forceAdaptiveThinking` 都是 RC8 原生字段，因此自适应思考和前端推理等级选择仍然有效。卸载后只会失去 Claude 专用的添加／编辑界面、模型探测和参数预填。
 
 ## 许可证
 

package/cordis.patch.yml

@@ -1,14 +1,5 @@
 - insert:
     - id: claude-provider
       name: '@mofeng2223/dsh-claude-provider'
-      config:
-        debug: !!js process.env.DSH_CLAUDE_PROVIDER_DEBUG === '1'
-        effortMap:
-          minimal: low
-          low: low
-          medium: medium
-          high: high
-          xhigh: xhigh
-          max: max
 - id: ui-settings-models
   disabled: true