MEMZ-JZY/DSH-Claude-Style-Reasoning-Slider0

dsh-client-ui-effort-slider

用于 DeepSeek Harness 模型选择的动画推理级别滑块

AI 分析

核心用途是提供类似 Claude 风格的动画推理级别调节滑块。适合需要直观调整模型推理强度(如 Off、Low、Medium、High 等)并快速切换模型的 Web UI 用户。

套件
dsh-client-ui-effort-slider
版本
0.7.0
授權
MIT
最近更新
2026年8月16日

安裝

$npx -p @deepseek-ai/dsh dsh plugin --profile web add github:MEMZ-JZY/DSH-Claude-Style-Reasoning-Slider

Usage

  1. Click the model/effort trigger chip in the conversation input.
  2. In the popup menu, choose Model to switch models, or Effort to adjust the reasoning level of the current model.
  3. In the effort pane:
    • Drag or click the slider to pick Off, Low, Medium, High, Extra, or Max.
    • If a position is unsupported, the thumb stays where you clicked and the nearest supported level below is applied; a toast tells you what happened.
    • Click Default to remove reasoningEffort and let the provider decide.
    • Click any extra pill to apply a provider-specific strength that does not map to a slider position.
  4. When switching models, the plugin preserves the current effort level where supported; otherwise it downgrades to the nearest supported level below or falls back to Default, always with a toast.
  5. To enable the Liang Calibrator, click the 滑动变祖器 switch below the track. The portrait and stage-name suffixes turn on immediately; the state is remembered across sessions.
  6. To enable the Big Fat Fish thumb, open DSH Settings → General and turn on Big Fat Fish slider. The thumb changes immediately without a restart.