A host-level pipeline with a workbench, parallel development, QA repair loops and restart recovery. It writes project artifacts and a managed AGENTS.md block, and reserves implementation for stage agents.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-10-30
License
MIT
01What can it help you accomplish?
Orchestrate a multi-agent development pipeline from a requirement to acceptance.
PRD, technical plans, development, QA and acceptance with stage tasks, artifacts and conclusions.
DeepSeek Harness users running team-style software development.
02How to install into DeepSeek Harness
Prerequisites
- DeepSeek Harness >=0.2.1-alpha.2 <0.3.0; README baseline: 0.2.1-alpha.4.
- Node.js 22.18 or later and a DSH host environment.
Installation steps
- 01
For the web profile, run `dsh plugin --profile web add dsh-plugin-teamflow`.
$ dsh plugin --profile web add dsh-plugin-teamflow
- 02
For desktop, install dsh-plugin-teamflow through the app’s Plugins page. Web and desktop installations are separate.
- 03
Restart `dsh --profile web`, or restart the desktop app for a desktop installation.
$ dsh --profile web
Verify the integration
- The conversation header and left sidebar show the Team Workbench entry.
Rollback
- For web, run `dsh plugin --profile web remove dsh-plugin-teamflow`, then restart `dsh --profile web`. Uninstall does not automatically clean optional runtime or project data.
03DSH integration and capability boundaries
Host-level TeamFlow service and web workbench loaded through a DSH profile.
Staged multi-agent development
A software requirement→PRD, technical plan, development, QA and acceptance artifacts
Pipeline documents are written under docs/teamflow/. Existing AGENTS.md files receive a managed block if absent. The author describes one closing commit per run.Parallel development and QA returns
Tasks split from the architecture plan and QA defects→Default three-way development concurrency, maximum eight; P0–P2 defects return for repair and retest up to two rounds
Resume and report
An interrupted pipeline run→Continuation from the first unfinished stage and a completion summary sent to the initiating conversation
04Who is it for? When not to use it?
Good for
- DeepSeek Harness users running team-style software development.
Not for
- The current lower bound is dsh 0.2.1-alpha.2, within >=0.2.1-alpha.2 <0.3.0. Older static 0.2.0-rc.2 checks in the README do not establish current runtime support.
- Web and desktop use separate installations. For desktop, the README directs users to the in-app Plugins page and says the CLI rejects --profile desktop.
- After teamflow_start, the main conversation should not independently edit code or run verification; the pipeline’s stage agents handle implementation, QA and reports.
- The package requires a DeepSeek Harness host and cannot run independently.
05Compatibility, maintenance and safety notes
- The current lower bound is dsh 0.2.1-alpha.2, within >=0.2.1-alpha.2 <0.3.0. Older static 0.2.0-rc.2 checks in the README do not establish current runtime support.
- Web and desktop use separate installations. For desktop, the README directs users to the in-app Plugins page and says the CLI rejects --profile desktop.
- After teamflow_start, the main conversation should not independently edit code or run verification; the pipeline’s stage agents handle implementation, QA and reports.
- The package requires a DeepSeek Harness host and cannot run independently.
Public GitHub source; README declares MIT.
06Frequently asked questions
Which current DeepSeek Harness version does TeamFlow require?
The current range is >=0.2.1-alpha.2 <0.3.0, with a README baseline of 0.2.1-alpha.4. The old static rc.2 check does not override the raised minimum.
Does installing TeamFlow in the web profile also install it on desktop?
No. The profiles have separate dependencies. Install desktop through its Plugins page; the README says its CLI does not accept --profile desktop. Restart the chosen host after installation.
Can the main DeepSeek Harness conversation keep coding after teamflow_start?
The README says it should not edit code or run verification itself. The pipeline’s stage agents handle implementation, QA and reports.
What happens after an interrupted TeamFlow run?
It resumes from the first unfinished stage and reuses completed artifacts. Completion sends a summary back to the initiating conversation.
07Related DSH workflows
archify
by tt-a1i
Turn any idea, plan, or codebase into a beautiful interactive diagram. An agent skill for Claude Code, Codex, and more.
openviking
by volcengine
Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
nocobase
by nocobase
NocoBase is an open-source AI + no-code platform for building business systems fast. Instead of generating everything from scratch, AI works on top of production-proven infrastructure and a WYSIWYG no-code interface, so you get both speed and reliability.
learn-harness-engineering
by walkinglabs
Harness engineering beginner tutorial, from 0 to 1
08Data and sources
TeamFlow 团队研发流水线 —— DeepSeek Harness 可分发插件（`dsh plugin --profile web add` 安装）。
- 浏览器侧：会话头部「🏭 团队工作台」tab（会话内）+ **左侧边栏「团队工作台」图标**（全局面板，产品线视角）；
运行需 DeepSeek Harness（dsh）宿主环境；本包不发布也无法独立运行。
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-10-30. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.