DeepSeek Harness プラグイン — 199 ページ

Five-decimal cache-hit readout beside the chat stats line for the DeepSeek Harness Web GUI

3D printing & additive manufacturing capabilities for DeepSeek Harness: parametric STL generation, calibration G-code, STL/G-code analysis and toolpath rendering as model tools, plus slicing / modeling / analysis / diagnosis knowledge skills.

Design, observe, diagnose, and reuse evidence-backed multi-Agent work in DeepSeek Harness

DeepSeek Harness plugin status panel / profile ops console: current profile, bundle state, pending-restart detection (SSE), one-click restart, multi-profile sync install with dual-end (web+desktop) mode, supply-chain version preview, rollback/undo, update detection & align, health checks, profile diff, boot report, audit log, and agent tools. · DSH 插件状态面板 / profile 运维控制台：观察、诊断、变更管理与 Agent 化。

Cache-stable progressive tool discovery and dispatch for DeepSeek Harness

DeepSeek Harness bundle: 项目罗盘 / Project Compass — 读取现有项目、解析多语言代码、构建统一 IR 与依赖知识图谱，生成带证据引用的上手报告（ONBOARDING / ARCHITECTURE / MODULE_MAP / KEY_FLOWS / GETTING_STARTED），并提供项目级带引用问答。

Project-level persistent memory plugin for deepseek-harness — auditable Markdown memory files with load / upsert / finalize / forget tools.

DeepSeek Harness 输入栏提示词优化插件：6 种策略 × 3 种语言与自定义指令重写草稿，可选携带聊天上下文（含压缩摘要）、目标与任务清单，结果确认弹窗、最近 5 次历史与三级设置持久化。Input-bar prompt optimizer for DeepSeek Harness: rewrites drafts with 6 strategies × 3 languages plus custom instructions, optional chat context (with compaction summary), goal and todo injection, a confirm dialog, last-5 history and three-tier settings persistence.

Reusable prompt profiles with per-turn model selection for DeepSeek Harness

Prompt Studio plugin for DeepSeek Harness: compose runtime prompts, overrides, and request-local synthetic turns in one live editor

Local prompt-injection and secret-exfiltration firewall for DeepSeek Harness

Bridge a GitHub Copilot subscription into DeepSeek Harness (dsh) as an LLM provider.

DeepSeek Harness plugin: live account balance/usage panel for every configured LLM provider (draggable floating-ball widget, configurable refresh interval and balance thresholds)

DSH web plugin: a bottom-right bubble showing the live usage/quota of every configured LLM model provider (OpenCodeGo, OpenRouter, OpenAI, DeepSeek…), with a configurable update policy (on user request / every N steps / interval / window reset / hover), theme-aware styling, device-local times and zh/en i18n.

Adds opt-in image-capable sibling routes for text-only providers and converts image blocks into budgeted, locally enhanced OCR, color-statistics, pixel-scan, and metadata evidence before delegation.

DSH 插件：拼图模式——一个会话只绑一个拼图项目（主文档 + 若干模块文档），AI 主动提问把不确定项变成已定项，用五维（任务复杂度 / 可拓展性 / 维护系数 / 代码质量 / 可复用性）算项目健康性；主文档第五节 `## 工作流` 是**标准化流水线**（一条 = 一个 `### 名字` 块 + 有序步骤，≤5 条 / 每条 ≤12 步，可整条删除进归档并恢复或永久删除）。文档格式 v6，带 op:rebuild 迁移、op:settings 按会话开关与三种执行模式（只拼不写 / 写后再拼 / 边拼边写）。v0.19.5 把提问写深：每题 3–6 个真岔路、每个配一句取舍，并给反例与正例（原先只说「别只给两三个」）；v0.19.4 把上限放开到一轮 10 问、每题 10 个选项，两个数收进 constants.js 的 ASK_MAX_* 一处定义。完整功能、安装与逐版本说明见 README。

PWA plugin for DeepSeek Harness - adds offline support and install capability

DeepSeek Harness plugin (temporary, pre-upstream-fix): automatically injects a CPython tempfile shim (PYTHONPATH -> sitecustomize) into every confined shell command on the Windows sandbox, so python/pytest tempfile use works with zero extra tools, zero model-context overhead, and zero escalation.

将 Qoder（国内版 / 国际版）的模型接入 DeepSeek Harness — bring Qoder (CN & international) models into DeepSeek Harness.

Provider quota/balance corner panel for the dsh web surface: server-side credential proxies plus a config-driven page badge for DeepSeek balance and OpenCode Zen Go usage windows.

选区引用与锚点批注：在 DSH 会话中选中文字 → 批注 → 引用锚点 chip（点击跳回原文、悬停显示原文）→ 提交序列化为结构化引用块。

DSH web plugin: quote any selected fragment of a conversation into the composer as a Markdown blockquote — smart floating button beside the mouse that dodges other overlays.

Provider-aware reasoning effort settings and composer control for DeepSeek Harness.

DeepSeek Harness 统一推理等级插件：一个 llm-reasoning 设置项动态配置 DSH 所有模型的默认推理等级（思考等级）——自动为手写声明的模型补齐推理能力、按路由设置默认等级（仅当全部模型支持）、同步 DeepSeek 官方路由，改设置即生效。