DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

PluginsPresetsDocsNews

Community

Publish a pluginContactReport an issue

Resources

Plugin Hub on GitHubDeepSeek HarnessSystem statusPrivacy notice
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

Autopilot — DSH Plugin for DeepSeek Harness
← Plugins
A

dsh-autopilot

Autopilot

Turn approved tracker tickets into review-ready pull requests with DeepSeek Harness.

The plugin will be installed here. Keep web if you are unsure.

npx -y @deepseek-ai/dsh plugin --profile web add github:canhta/dsh-autopilot#84a79a89b794fbd5721ef14764afac4423a35b4f
READMECompatibilityVersions

Compatibility and provenance

Autopilot is published as dsh-autopilot and currently resolves to version 0.0.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated
9/11/2026

Versions

0.0.0stable
9/11/2026

Related plugins

Loading related plugins…

Latest
0.0.0
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
any
License
MIT
Source
github
GitHub
★ 0
Weekly downloads
0
Last push
9/14/2026
View source ↗
README badge

Click the badge to copy Markdown for your README.

Do you maintain this Plugin?Claim benefit · Priority security scan

Verify the GitHub repository declared in package.json to manage this listing. After you claim it, Hub will prioritize a security scan of the current version and publish the result when it passes.

Claim this Plugin →
Report an issue
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in

Related plugins

More verified plugins in productivity-workflow.

Acp App@deepseek-ai/dsh-acp-appThe dsh ACP profile bundle: automation-only JSON-RPC stdio and process lifecycle over dsh-baseClient Ui Task Board@linxin666/dsh-client-ui-task-boardHost-authoritative task board for the DSH Web GUI with real session execution, Host cron scheduling, and optional cross-platform idle-sleep protection; mounted without DSH source changes.Web All@linxin666/dsh-web-allDSH Web UI 全家桶聚合插件:一键安装全部功能插件(task-board / git-graph / pet / remote-web-ui / web-ui-settings / skin-center / community-plugins / compat shim)。compat 桥接层已并入本包(src/client),无需独立 compat npm 包。Agent Teams@nanmicoder/dsh-agent-teamsAgentTeams for DeepSeek Harness: multi-agent team collaboration (captain, members, tasks with dependencies, messaging) driven by natural language, with a tree monitor in the web GUI

README

dsh-autopilot — Turn tickets into pull requests. Powered by DeepSeek Harness.

dsh-autopilot

dsh-autopilot is an open-source plugin for DeepSeek Harness (DSH). Its goal: move approved tickets from your backlog to review-ready pull requests on your VPS, without supervising every agent turn.

You decide which tickets are ready, when the agent can work and how much it can spend. Autopilot coordinates the work; your repository defines how code is built and checked; people keep control of blockers, review and merge.

Pre-release · bootstrap available. The repository now ships a loadable DSH bundle and its build/test/package workflow. Ticket intake, execution, provider connections and the Web UI are still planned capabilities, not working features. Follow each implementation slice in GitHub Issues.

Product direction · Installation · Local development · Contributing · Documentation

Product direction

  • Work from your existing backlog. Select tickets by a configured label and approved brief. Use Jira or Linear with GitHub or Bitbucket; add providers through Cordis plugins.
  • Control when work starts. Set a schedule, execution limits and spending policy. Eligible work waits in a durable queue until it can run.
  • Pause without losing the workspace. Retain the run, DSH Session and Git worktree so interrupted work can continue.
  • Resolve blockers where the task lives. Receive questions in ticket comments. A human answers and explicitly marks the ticket ready to continue.
  • Stay informed. Configure ticket comments, webhook or ntfy notifications. Manage runs, schedules, budgets and retained worktrees inside DSH Web.
  • Receive a PR ready for human review. The agent follows the target repository's instructions and local checks. Autopilot stops at PR creation; it does not merge or manage remote CI.

How it works

flowchart LR
    accTitle: From approved ticket to review-ready pull request
    accDescr: Approved tickets enter the Autopilot queue. DSH executes eligible work under schedule, capacity and budget limits. Verified local work becomes a review-ready pull request. Notifications are delivered independently after PR creation.

    ticket["Ready ticket<br/>Jira / Linear"]
    queue["Autopilot queue<br/>Schedule + limits"]
    agent["DSH agent<br/>Code + local checks"]
    pr["Review-ready PR<br/>GitHub / Bitbucket"]
    notify["Notify<br/>Configured channels"]

    ticket --> queue --> agent --> pr
    pr -.-> notify

Autopilot selects tickets with the configured ready label and approved Agent Brief, then dispatches eligible work when the schedule, capacity and budget allow. DSH follows the repository's rules in a Git worktree. Creating the verified PR completes the run; notification delivery is independent. Humans own review, merge and marking the ticket Done.

  • Blocked: questions go into ticket comments. A human must reply and mark the ticket ready before work can continue.
  • Paused: keep the Session and worktree. Resume rechecks eligibility and execution limits; an operator pause requires explicit resume.

See run lifecycle for exact pause, recovery and publication rules.

This is an issue-driven AI Development Lifecycle (AIDLC) flow for one project on one VPS. Autopilot adds coordination to DSH; it does not replace the harness's agent runtime, tools or Web application. See product responsibilities for the precise division of ownership.

Installation

The bootstrap can be installed from a locally packed artifact. It does not yet automate tickets, so no Jira, GitHub or model credential is required for this verification path.

Prerequisites: Git, Node.js 24 or newer, and pnpm. DSH itself is invoked from the npm latest tag.

git clone https://github.com/canhta/dsh-autopilot.git
cd dsh-autopilot
pnpm install
mkdir -p .artifacts
pnpm pack --pack-destination .artifacts

npx --yes @deepseek-ai/dsh@latest --profile autopilot --from-default-profile web --dump-config
npx --yes @deepseek-ai/dsh@latest plugin --profile autopilot add ./.artifacts/dsh-autopilot-0.0.0.tgz
npx --yes @deepseek-ai/dsh@latest --profile autopilot --dump-config
npx --yes @deepseek-ai/dsh@latest --profile autopilot --no-open

The third DSH command must show a dsh-autopilot bundle layer containing the autopilot row. DSH prints the local Web URL when the final command boots. Omit DSH_HOME to use DSH's default profile location, or set it to an operator-owned directory before all four DSH commands to isolate the installation.

The future connection flow will live in DSH Web Settings. Editing .env is not the intended onboarding path; see operations and configuration for the planned credential ownership model.

Local development

Install the checkout and run all local gates:

git clone https://github.com/canhta/dsh-autopilot.git
cd dsh-autopilot
pnpm install
pnpm run hooks:install
pnpm run check
mkdir -p .artifacts
pnpm pack --pack-destination .artifacts
pnpm run verify:package

pnpm run check runs Biome, strict type checking, the test suite and a clean production build. pnpm pack repeats the static gates and builds the distributable tarball. pnpm run verify:package installs that artifact into a disposable DSH profile, boots its built Host entry, and proves that package resolution rejects a missing Host entry with the expected diagnostic. Start with the contributor guide, then use the documentation map to find the relevant specification and source research.

Contributing

Contributions are welcome: clarify a requirement, verify a DSH integration against source, improve onboarding, or implement an agreed issue. Find or open a GitHub issue before substantial work so scope and dependencies are clear.

See Contributing for the local workflow, engineering rules and what to include in a pull request. Keep implementation progress and review evidence on GitHub, not in the specifications.

License

MIT. Community project; not an official DeepSeek product.