dsh-reasoning-effort-new：为 DSH 任意模型配置推理强度
dsh-reasoning-effort-new 由 Ayowait 维护，为 DeepSeek Harness 中已接入的任意模型补齐推理强度 settings，支持 Provider 默认、模型覆盖与 wire 映射，并从 llm-pi-ai 自动发现 Provider。Web 端提供五档离散滑块与按 Provider 分组的模型选择器，最高档有紫色动效。安装：dsh plugin --profile web add github:Ayowait/dsh-reasoning-effort-new 后执行 dsh web。默认不覆盖已有 reasoningEfforts；特殊网关可在 Cordis patch 中配置。MIT 许可证。
Read Moredsh-codex-power-slider：DSH Web 上的 Codex 风格推理强度滑杆
dsh-codex-power-slider 是 NNNOOOOMII 维护的 DSH Web 第三方客户端插件，在 composer 旁接管模型入口，用 Codex 风格卡片滑杆读写会话真实推理强度，并支持分组模型列表与 Ultra 档动效。安装命令为 dsh plugin --profile desktop add dsh-codex-power-slider。仅支持 web 平台，subagent 路由会话中会禁用滑杆。MIT 开源，零构建零 Node 依赖。
Read Moredsh-reasoning-slider：在 Harness 里用滑块切换四档推理强度
dsh-reasoning-slider 是 Xylocarpro 维护的 DeepSeek Harness 客户端 UI 插件，在模型名称旁提供类 Codex 的四档推理滑块（关/轻度/高/Ultra，对应 off/low/high/max），并集成模型搜索与档位能力校验。当前版本 0.1.40 在 Windows Harness 0.2.0-rc.2 上验证，通过发行 ZIP 与 install.ps1 安装到 desktop profile。本文基于官方 README 介绍功能、安装命令、典型用法及插件权限与兼容性注意点。
Read MoreDeepSeek Harness 插件：按模型配置推理强度
插件 `dsh-model-reasoning-levels` 为 DeepSeek Harness 的原生模型设置页面补全了 `reasoningEfforts`（推理强度）的行内编辑功能。它通过挂载编辑器并调用 `settings.mutate`，将配置写入 `cordis.patch.yml`，支持按模型级别调整推理强度。插件提供中英文界面，强制执行严格的验证规则，确保配置的有效性，并允许用户自定义推理能力集。
Read MoreDSH Model Controller: Switch Models, Adjust Reasoning Intensity, Favorite Models, and Reorder Providers
The DeepSeek Harness plugin (DSH) adopts a plugin-based architecture and takes over the model slot to enhance model selection. Its core features include searching, filtering by provider, favoriting frequently used models, and adjusting thinking intensity. The thinking intensity level has per-model memory and automatically restores the corresponding level when switching models. For installation, download the `.tgz` file and install it via the host; the plugin will automatically take over the official selector. It supports integration with Magpie and has fallback interface configuration capability. Data is stored locally on the client; preferences are retained after uninstalling and reinstalling. It is compatible with DSH 0.1.7-rc.2 and Cordis 4.0.4.
Read Moredsh-effort-ultra: DeepSeek Harness Native Reasoning Effort Control
`dsh-effort-ultra` is a reasoning-effort control plugin designed exclusively for DeepSeek Harness (DSH). By calling the official DSH session service, it resolves the lack of native interface controls and conflicts with third-party controls, offering six adjustment levels (from Off to Ultra) and a blue-to-purple gradient interface. The plugin supports continuous sliding, explicit menus, model-switch synchronization, and theme adaptation. It directly reads the `reasoning` declaration in the official DSH directory to ensure accurate levels. Easy to install and highly compatible, it is an ideal tool for developers who want precise control over model outputs.
Read Moredsh-task-effort: Automatically Adjust Model Reasoning Effort Based on Task
This plugin is designed for DeepSeek Harness (DSH) agent development and automatically adjusts the model’s `reasoningEffort` parameter by intelligently recognizing task content. It assesses task complexity based on keywords and message length, and supports automatic effort capping during peak hours to reduce costs. The plugin provides manual fallback mechanisms (GUI or `[effort=...]` tags) and automatic escalation on errors. By injecting system reminders and capsule displays, it allows users to clearly see the current effort level and the rationale behind it, effectively controlling token consumption while preserving reasoning quality.
Read Moredsh-reasoning-effort-slider: DSH Reasoning Effort Slider Plugin
The DSH reasoning-effort slider plugin, dsh-reasoning-effort-slider, is designed to simplify configuration management for model reasoning effort. It provides a 7-level visual slider, automatic configuration writing, self-healing for anomalies, and visual feedback such as chibi animations. After terminal-based installation, users can apply a level instantly through hot listening without restarting. Additionally, the plugin includes a built-in knowledge base for level conversion and compatibility fixes for pi-ai models, greatly lowering the barrier for DSH users to adjust reasoning effort.
Read Moredsh-reasoning-effort-slider: DeepSeek Harness Reasoning Effort Slider Plugin
`dsh-reasoning-effort-slider` is a plugin for DeepSeek Harness, designed to address the issue where reasoning effort is silently turned off when switching models in the native logic. The plugin injects a slider into the chat bar, supports multiple models including Bailian and DeepSeek, and allows independent adjustment of reasoning effort per conversation. Its core features include smart UI display (showing only supported levels), automatic memory of each model’s configuration, and automatic inheritance or graceful downgrading of the remembered effort level when switching models. This mechanism prevents thinking capability from being silently disabled and supports installation with a single command.
Read Moredsh-ui-skin-switcher: DeepSeek Harness Model and Reasoning Effort Switcher
`dsh-ui-skin-switcher` is a plugin built for the DeepSeek Harness Web interface. It provides a compact Claude-style switcher next to the input box, supports selecting models grouped by vendor, and offers an animated thinking-intensity slider for models that support reasoning. The plugin includes per-model memory and automatic theme adaptation. It is ready to use after installation and restart, but note that it only supports DSH 0.1.5+, and the thinking intensity must be configured manually—an incorrect value will silently fail.
Read Moreluxueliu-reasoning-efforts: Enable gateway models in DSH to select reasoning intensity; one-click level adjustment after upgrade
## Introduction In DSH, the reasoning effort selector only takes effect for ds-series models. If a model connects via a local LLM gateway with manual declarations—for instance, aliases such as gemini3.x, grok, kimi-*, or glm5.x—it neither displays the effort selector nor allows switching the thinking intensity during the conversation. Previously, users could only manually supplement effort declarations for individual models in `~/.dsh/settings.yaml`. However, after a DSH upgrade or when the interface rewrites the settings, these configurations would drift, necessitating repeated rewriting. Below, I introduce the luxueliu-reasoning-efforts plugin. It consolidates the tier configuration into a single source of truth, providing checks and...
Read Morehytime/dsh-thinking-effort: Supplementing the Thinking Intensity Settings for DSH's llm-pi-ai Model
`@hytime/dsh-thinking-effort` is a DSH plugin designed to supplement default configurations for third-party `llm-pi-ai` models that lack declared thinking effort levels. Its core functionalities include adding default levels of `off`, `high`, and `max` without overriding custom values; supporting the selection of thinking effort levels per model within DSH settings and mapping DSH levels to corresponding gateway values (e.g., `high` mapping to `ultra`); setting default thinking intensity for sub-agents, which only applies when not explicitly specified in a request; supporting multilingual interfaces in Chinese, English, Japanese, and Korean (with Japanese/Korean requiring DSH core support); and displaying a version watermark on the settings page. Users can manage the plugin via the DSH CLI command `dsh plugin --pr
Read MoreAdd Visual, Reasoning Tier and Token Capacity for Custom Models in DeepSeek Harness via better-model-provider
### 正式翻译： The official "Models" page of DeepSeek Harness can manage providers and API keys, but in the past, adjusting the inference strength levels, visual input capabilities, and token capacity for custom models could only be done by manually editing the YAML files. The MIT-licensed plugin `better-model-provider` maintained by sanshanya added a "Model Capabilities" section on the settings page, which configures these native capabilities based on the model's declared specifications, without modifying the Harness runtime or touching credentials. This article verifies the installation commands, usage methods, and compatibility boundaries based on the community directory page and the GitHub README.
Read MoreAdd Model and Inference Strength Sliders to DeepSeek Harness Using dsh-reasoning-effort
# 修正：原文中“大肥鱼滑块”应为“大模型推理强度滑块”，以下是准确翻译： HanaAyane-maintained dsh-reasoning-effort brings Codex-style models and reasoning effort sliders to the DeepSeek Harness Web interface, with the gear positions reading model directories and syncing with `/model`. This article verifies the installation commands, the read-only YAML guidelines for custom providers, as well as the notes on the large model reasoning effort slider, permissions and licenses based on the directory page, GitHub README and v0.6.0 version.
Read More