DSH Plugin HubDSH 插件市场
linxuhao/AItelier 预览 preview

linxuhao/AItelier

Plugin插件 ⭐ 3 MIT Subagents & Orchestration子代理与编排Prompts & Skills提示词与技能

Structured-but-dynamic subagent workflows for AI agents — your agent delegates work to deterministic, fully-audited pipelines it can generate, run, and edit over MCP.

catalog descriptioncatalog 简介 / catalog description：AI-Atelier, the all in one personal "atelier" (means handcraft studio in french) that can adapt to your need.

Project Overview项目介绍

Edited by Zhiwei Agent (zhiwei)由知微Agent(zhiwei)编辑

AItelier is an open-source MCP-compatible multi-agent workflow engine that works with any MCP-enabled agent including DSH. It makes multi-agent AI pipelines deterministic and fully auditable, and exposes all functionality over MCP so external agents can delegate bulk work to low-cost deterministic pipelines, only making decisions at designated checkpoints. It is built on top of the open-source SkillFlow engine, which is available on PyPI as skillflow-py. To install AItelier locally, you can clone the repository and run pip install -e . from the project root, or rebuild the provided Docker image if you prefer containerized deployment. It currently ships a production-ready software delivery pipeline, with a no-code visual workflow platform planned for future releases.

AItelier provides a persistent State DAG for managing long-lived goals, revised acceptance contracts, dependencies, and audit evidence. SkillFlow handles workflow steps, loops, retries, and checkpoints, while the State DAG tracks project status and delivery events. It exposes MCP tools for reading, writing, and waiting for state changes, as well as tools for managing driver notes that let multiple project directors work in parallel without conflicting revisions. It also supports external verification of nodes without creating a new SkillFlow run, so directors, subagents, CI tools, or proof-checking harness can submit scoped artifacts and evidence for verification. This tool is ideal for AI developers and engineering teams that need auditable multi-agent collaboration pipelines for production use.

AItelier is released under the permissive MIT license, and requires Python 3.12 or newer to run. The project structure separates core functionality into modules for configurations, templates, tools, core logic, API endpoints, a web frontend, and a CLI. The web frontend is built with Svelte 5 and Vite as a single-page application, supports 8 languages with live switching, and is served by the FastAPI backend. The project maintains nearly 950 unit and integration tests to validate functionality, and you can run the full test suite with the command pytest tests/ -v after installation. To build the frontend locally, navigate to the web directory and run npm install && npm run build to compile the bundle for serving.

AItelier是一个兼容MCP协议的多智能体工作流引擎，支持所有支持MCP的AI代理，包括DeepSeek Harness（DSH）。它的核心能力是让多智能体AI流水线具备确定性和完整可审计性，可通过MCP接口供外部代理调用，主代理可将批量工作委托给低成本的确定性流水线，仅在检查点做决策。它基于开源的SkillFlow引擎构建，目前已提供成熟的软件交付流水线，未来规划推出无代码工作流平台。

它提供持久化的状态有向无环图（State DAG），用于管理长期目标、验收契约、依赖项和审计证据，支持工作流步骤、循环、重试和检查点管理。用户可通过MCP工具调用状态图读写、等待状态变更、管理项目笔记等功能，支持多个主管并行管理项目而不冲突。这套工具适合需要构建可审计多智能体协作流水线的AI开发者、团队和合规要求较高的项目。

AItelier采用MIT许可证开源，依赖Python 3.12+，项目结构清晰，分为配置、模板、工具、核心、API、前端Web和CLI模块。前端使用Svelte 5 + Vite构建SPA，支持8种语言切换，CLI提供富文本TUI面板。可通过pip install -e .安装，也支持Docker镜像构建，提供近千个单元和集成测试保障稳定性。

View on GitHub在 GitHub 查看

Recommended强烈推荐 Send to your DSH发给你的 DSH 安装 
请帮我安装这个 DSH 插件。安装前先完成【兼容性检查 + 安全性检查】，检查通过再动手。

插件：AItelier（linxuhao/AItelier）
仓库：https://github.com/linxuhao/AItelier
本站详情页：https://www.yhbd.top/plugins/linxuhao-aitelier/
本站登记：类型 plugin · 归类 多平台兼容工具（非 DSH 原生） · 许可证 MIT · ⭐ 3 · 最近提交 2026-10-01 · 主语言 Python · 未检测到 DSH 插件清单

按下面顺序执行，每步先把结论告诉我，再进入下一步：

【1 兼容性检查】
① 我这边：DSH 版本、Node 版本、操作系统、当前 profile（web / desktop）。
② 读它的 README、package.json、插件 manifest，列出它要求的 DSH 版本 / Node 版本 / 操作系统 / 外部依赖 / 需要另外先装的运行时。
③ 逐条比对，结论只写「满足 / 不满足 / 未知」三种；不满足的给出可行替代方案。
④ 检查是否和我已装的插件冲突：命令名重复、skill / tool 重名、端口占用、重复注册的 MCP server。

【2 安全性检查】
① 仓库可信度：和上面「本站登记」是否一致；star / fork 数、创建时间、最近提交，是否归档或长期停更。
② 安装脚本：逐行看 package.json 的 preinstall / install / postinstall，以及 install.sh、setup.ps1 之类脚本。出现 curl|bash、下载后直接执行、混淆代码、访问与插件功能无关的域名，立刻停下来告诉我，不要继续装。
③ 依赖：列出新增依赖，标出无人维护、或与知名包拼写近似的可疑包（typosquatting）。
④ 权限与副作用：它会读写哪些目录、访问哪些域名、需要哪些 DSH 权限（filesystem / network / shell / clipboard 等），以及怎么卸载和回滚。
⑤ 如果它要求 sudo / 管理员权限，或权限明显超出功能所需，先停下来问我。

【3 安装】
上面两步没有「不满足」和「高危项」时才执行；用官方推荐方式安装，不要自行提权。

【4 汇报】
用表格输出：检查项 / 结论 / 依据 / 是否需要我决策。拿不准的一律写「未知」并说明要我怎么确认——不要猜，也不要替我决定。

Send this message to DSH in your current session: it verifies compatibility and security first (answering met / not met / unknown item by item) and only installs once everything checks out — it will stop and ask you if it finds a high-risk item. The box scrolls; the copy is the full prompt. CLI install commands may not be accurate across systems, so DSH is the safer route.把上面这条消息直接发给当前会话里的 DSH：它会先核对兼容性与安全性（逐条给「满足 / 不满足 / 未知」），确认没问题再安装，有高危项会停下来问你。框内可滚动，复制到的是完整提示词；安装命令不一定准确，发给 DSH 更稳。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 2 warnings2 项注意
  • Only 3 stars - very few users, little community feedback星标只有 3，几乎没人在用，遇到问题缺少社区反馈
  • No DSH plugin manifest detected - it may only carry the dsh-plugin topic, so the install method must be confirmed on the spot未检测到 DSH 插件清单：可能只是打了 dsh-plugin 话题，安装方式要现场确认
  • Not DSH-native: a multi-platform tool that may require Node / Electron or another runtime first非 DSH 原生，是多平台兼容工具：可能要先装 Node / Electron 等运行时
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时（Electron / Python / Docker 等）是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突：命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致，是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需（filesystem / network / shell / clipboard）
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限，以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照，不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装（适合开发者）

CLI Install命令行安装 

dsh plugin --profile web add github:linxuhao/AItelier

把 linxuhao/AItelier 加入你的 DSH 配置（web profile）即可启用。

READMEREADME

32.8 KB · 缓存于 2026-09-17 19:38:16 · 32.8 KB

AItelier

Structured-but-dynamic subagent workflows for AI agents — your agent delegates work to deterministic, fully-audited pipelines it can generate, run, and edit over MCP.

License Python Engine: SkillFlow

AItelier makes multi-agent AI pipelines deterministic and fully auditable — define a pipeline (or have your agent generate one), run it, and inspect why it did everything it did. The whole surface is exposed over MCP, so any MCP-speaking agent can use AItelier as its workflow engine: delegate bulk work to cheap, deterministic pipelines and only decide at checkpoints (see Use AItelier from another agent). Under it all is an open engine (SkillFlow, MIT, on PyPI as skillflow-py) plus a flagship software-delivery pipeline; the broader no-code workflow platform is on the roadmap.

Persistent project state, separate from workflow execution

AItelier also provides a State DAG for long-lived goals, revisioned acceptance contracts, dependencies and evidence. SkillFlow continues to own workflow steps, loops, retries and checkpoints; the driver chooses a ready goal and a workflow. A completed workflow produces a candidate, not a verified product capability.

The MCP/internal-driver tools state_graph_help, state_graph_read and state_graph_write expose the same typed contracts as /api/state. State data reads require writer authorization. Existing DPE pipelines and task/project UI remain compatible; legacy tasks are imported only explicitly and never inherit verified status. See architecture, usage, trust boundaries and rollout and the offline real-engine demonstration. The authenticated send_director_message, list_director_messages, acknowledge_director_message, and resolve_director_message State actions provide project inboxes with durable delivery events; REST exposes the same closed v2 contract at /api/state/director-messages/<action>. The v2 lifecycle distinguishes explicit-inbox transient deliveries from standing guidance that remains in a bounded, redacted PostCompact recovery projection until resolved. Existing v1 rows migrate as transient without losing their message, delivery, event or idempotency audit. The State Project frontend and migration preparation guide covers project DAG browsing, exact-run graph versions, protected historical references and held shadow migration rehearsal. The project-first dashboard guide covers the default State DAG workspace, separate Runs/Pipelines navigation, readable node badges and actual run history.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要，完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

渲染基于仓库 linxuhao/AItelier 的 README.md。 在 GitHub 查看原文 →

Related相关推荐

← 上一个 Prev dsh-opencode-go-usage 下一个 Next dsh-llm-rate-limiter →