v0.2.7 · Apache-2.0
Prompt enhancement for the dsh web GUI: rewrite the composer draft into a well-structured prompt through the harness LLM service, preview the before/after, then fill back, copy, or undo. Original draft is always preserved.
v0.0.7 · BSD-3-Clause
Lark/Feishu IM bot channel for DeepSeek Harness | 飞书 DeepSeek Harness (DSH)插件
v0.0.0 · MIT
A focused end-user QA surface backed by native DeepSeek Harness sessions
v0.1.0 · MIT
DeepSeek Harness plugin: give your agent a browser with a persistent identity - engine-level fingerprint spoofing, unlimited free local profiles, Android device emulation, passkeys that survive, and residential proxy egress.
v0.18.0 · SEE LICENSE IN LICENSE
Enterprise-grade, local-first Agent Workbench for people and agent teams. A unified multi-engine workspace for Codex Harness, DeepSeek Harness, and OpenCode, with unified plugins and Skills, multi-agent projects and tasks, and editable code, documents, presentations, design, and video.
v0.3.0 · Apache-2.0
Native Rapid-MLX provider for DeepSeek Harness (dsh) — dsh reads model facts from the server instead of your settings.yaml.
— · Apache-2.0
Self-evolving memory OS for LLM & AI Agents: ultra-persistent memory, hybrid-retrieval, and cross-task skill reuse, with 35.24% token savings and DeepSeek Harness support.
v0.3.3 · MIT
把 dsh agent 接入微信、飞书等 20+ 聊天平台的聚合网关插件 | Aggregate IM gateway for DeepSeek Harness (dsh): connect your agents to WeChat, Feishu, Telegram, Discord & 20+ chat platforms
v0.7.2 · MIT
DeepSeek Harness (dsh) plugin: prompt your home servers into existence through home-hosted — Gitea, Jellyfin, anything — and manage the panel, the entries and boot autostart from dsh.
— · MIT
本地私有、开源的自进化跨平台 AI 内容发现 Agent:先理解你,再主动从 B站、小红书、抖音、YouTube、X、知乎、Reddit、微博等平台与开放 Web 寻找内容。(支持 deepseek harness 插件) | Local-first open-source cross-platform AI content discovery agent: understands you, then proactively finds content across Bilibili, Xiaohongshu, Douyin, YouTube, X, Zhihu, Reddit, Weibo and the open web.(support deepseek harness plugin)
— · MIT
DSH Plugin Radar — open-source ecosystem radar for DeepSeek Harness plugins: continuous discovery (21k+ candidates), k8s runtime validation (13k+ tests), 15-min snapshots; the catalog is a generated artifact — 开源 DSH 插件生态雷达:持续发现 2.1 万+ 候选、k8s 运行级实测 1.3 万+、15 分钟快照;插件目录为自动生成的产物
v0.1.12 · MIT
DeepSeek 写网文/小说的工作流插件:DeepSeek Harness 社区插件,内置小说、短剧、游戏、视频解说四个工作台 | Community DeepSeek Harness plugin with novel, short-drama, game and video-recap workbenches.
— · MIT
🖥️ Agent-project workbench for DeepSeek Harness — sidebar app drawer + dockable split workspace + a live control room watching every project.
v0.4.1 · Apache-2.0
TypeSafe Jev for DeepSeek Harness, the Model Context Protocol, and plain Node: typed judgments instead of prose, offline by default.
— · MIT
DeepSeek Harness Mobile 是一个面向 DeepSeek Harness 的原生 iOS 客户端。它通过 dsh-plugin-mobile-gateway 与 Harness 建立 WebSocket 连接,将工作区、会话、实时回复和 Agent 执行轨迹带到 iPhone,同时延续 DeepSeek WebUI 克制、清晰的视觉语言
v1.2.0 · no license
中华传统智慧(玄枢)AI Agent 技能包的 DeepSeek Harness(dsh)Cordis 插件:八字/紫微/六爻/梅花/奇门/风水/五运六气/体质全融合,本地确定性引擎 + 可视化 Dashboard,一行 dsh plugin add 安装。
— · MIT
DSH plugin — free OpenCode Zen models for DeepSeek Harness (DSH). Free LLM API, no API key needed. 在 DSH 中使用 OpenCode Zen 免费模型,无需 API key
v2.0.20 · MIT
Reflect2Evolve memory plugin: layered L1/L2/L3 memory, reflection-weighted value backprop, cross-task policy induction, skill crystallization, and three-tier retrieval for OpenClaw, Hermes Agent, and DeepSeek Harness.
— · MIT
Harness the power of local LLMs with this TUI MCP Client for Ollama. Featuring all core MCP primitives (tools, prompts, resources), agent mode, multi-server, model switching, streaming responses, human-in-the-loop, thinking mode, model params config, system prompts, and saved preferences.
v0.1.0 · MIT
在 DeepSeek Harness / Claude Code / Cursor / Codex / Gemini CLI 里直接搜索 20 个中国开放平台的 65,600+ 篇 API 文档;零配置,支持 Skill 与 DSH 原生插件。
v0.2.19 · GPL-3.0
AI 驱动的法律行业案件管理与 Agent 工具:诉讼案件 / 非诉项目 / 任务中心 / 期限引擎 / 技能与 MCP,Agent 预设直接办案,运行于 DeepSeek Harness。
v0.4.67 · MIT
T Expert — a 623-expert, 22-division roster with full Chinese translations, as a standalone DeepSeek Harness plugin.
v0.2.26 · MIT
Active Context Pruning (ACP) for the DeepSeek Harness — model-driven context management as a CompactionEngine backend.
— · MIT
Best ready-to-run DeepSeek Harness distribution: DSH desktop GUI, DSH terminal TUI, and harness integration.