At a glance
What it does
Schedule standalone coding tasks in isolated DeepSeek Harness sessions.
Web Profile
DSH peers: >=0.1.0-rc.5 <0.2.0; Cordis: >=4.0.1 <5.0.0
Evidence-verified
Checked Sep 6, 2026, 1:50 PM UTC
Code-evidenced contributions
What it adds to DSH
Lets an agent create and manage scheduled coding-task rules through the Host approval flow.
Mechanism evidence ↗Adds Scheduled sidebar and Settings pages for task rules and run history.
Mechanism evidence ↗Before you choose it
A community-maintained DeepSeek Harness plugin for creating, pausing, resuming, running, and deleting scheduled coding-task rules. It adds a Scheduled sidebar plus Settings pages for rules and durable run history, and can create rules from chat through the Host approval flow.
Best for
DeepSeek Harness Web users who want recurring or one-off coding jobs managed as separate sessions.
Common tasks
- Run a coding prompt once, at an interval, hourly, daily, weekly, monthly, or every N days.
- Choose the workspace, model, skills, and Host permission preset for a scheduled run.
- Pause, resume, trigger, or delete a rule and filter its recorded run history.
- Describe a schedule in chat and confirm its creation with the Host approval card.
Permissions and data
Scheduled runs use the Host's available tools, permission presets, and approval checks.
Permissions- Users choose from Host-provided permission presets, including any custom presets.
- A full-access preset follows the Host's risk confirmation; unattended runs are described as fail-closed `never`.
- Stores durable task definitions and run-history statuses.
- Each run uses its saved prompt, workspace, model, and permission boundary; source-chat history is not inherited.
- Requires a working DeepSeek Harness Web installation and its configured model/provider services.
- No plugin-specific credential requirement is declared; any Host or model credentials are not described in the supplied evidence.
Limitations
- It is a community-maintained plugin, not an official DeepSeek AI product.
- One active run is allowed per rule; overlapping occurrences are recorded as skipped.
- There is no automatic retry after a started run.
- A Host restart marks leftover queued or running records as failed with `host_interrupted`.
- Runtime behavior and installation were not executed during this curation.
What DSHub checked
- Pinned Git source and DSH bundle structure were verified.
- The npm registry identity for @michengai/dsh-automation version 0.1.32 was verified.
- The manifest declares Apache-2.0 and Node.js ^22.19.0 or >=24.0.0.
What DSHub did not check
- Successful installation in a DSH profile was not performed.
- The repository's tests, build, UI, scheduling, and permission behavior were not executed.
- Published npm package contents were not audited.
Pinned install
Install DSH Automation
This plugin bundle does not have a DSH Plugin install action. Use its source documentation for the delivery method.
Maintainer source
Project README
DSH Automation
Run standalone coding tasks on a schedule in DeepSeek Harness
简体中文 · Changelog · Apache-2.0
</div>DSH Automation is a community-maintained DeepSeek Harness (DSH) plugin, not an official DeepSeek AI product.
Features
- Manage scheduled tasks from Settings → Scheduled Tasks.
- Create, pause, resume, run now, and delete rules from the Web UI or Agent tools.
- Start each occurrence in a fresh root Agent and Session. Source-chat history is not inherited.
- Support once, interval, hourly, daily, weekly, monthly, and custom-every-N-days schedules.
- Pick workspace, model, skills, and any permission preset exposed by the Host.
- Create from chat: describe the schedule in any conversation, then confirm with the official approval card.
- Keep durable run history:
queued,running,succeeded,failed,skipped,cancelled. - Add a sidebar Scheduled tab. Folders are task names and child sessions are run times. On stock DSH it wraps the official workspace tree and does not depend on
dsh-codex-ui.
Interface
Scheduled tasks live in the workspace Scheduled tab, next to Tasks and Channels:

Open Settings → Scheduled Tasks to search, create, pause, and inspect rules:

Describe the job in chat. The agent calls automation_create and asks through the official approval card:


After approval, the rule is saved and summarized in the conversation:

Run history stays in Settings and can be filtered by day, week, month, task, or status:

DSH product ecosystem
This product can be installed independently or used through the desktop app or Web suite. They share the same DSH core but serve different ways of working; on stock DSH, this product does not depend on Codex UI:
| Product | Relationship to this product |
|---|---|
| DeepSeek Harness | The host runtime that provides models, sessions, tools, and the plugin system |
| DSH Codex Desktop | A ready-to-install desktop product with this product and the other five feature products built in |
| Six feature products | Codex UI · IM Connect · Automation · Skills Manager · Archive Manager · Agency Agents |
Prerequisites
- A working DeepSeek Harness Web installation with
dshavailable in PowerShell. - Examples use the
webprofile; replace it with the target profile. - Source installation and development require Node.js 22.19+. npm installation does not require running
npm installin an arbitrary directory.
Installation
dsh plugin add forwards to pnpm add in the profile directory. If you omit a version or the official registry, a local mirror may leave you on an old build.
Install from npm
[Console]::OutputEncoding = [System.Text.Encoding]::UTF8
$OutputEncoding = [System.Text.Encoding]::UTF8
dsh plugin --profile web add @michengai/dsh-automation@latest --registry=https://registry.npmjs.org/
dsh --profile web --dump-config
Restart DSH Web and hard-refresh the browser. Pin a version with @0.1.5 instead of @latest when needed.
Install from source
Use this for debugging or unpublished changes. The cloned directory becomes the plugin source path:
[Console]::OutputEncoding = [System.Text.Encoding]::UTF8
$OutputEncoding = [System.Text.Encoding]::UTF8
Set-Location D:\Repository\deepseek-harness-plugin
git clone https://github.com/MichengAI/dsh-automation.git
Set-Location .\dsh-automation
pnpm install
pnpm test
pnpm build
dsh plugin --profile web add .
dsh --profile web --dump-config
Restart DSH Web and hard-refresh the browser. Local installation reads and applies cordis.patch.yml; do not copy lib files manually.
Usage
Open Settings → Scheduled Tasks, then use the panel as follows:
| Goal | Action | Scope |
|---|---|---|
| Create a rule | Select New scheduled task, then set name, schedule, prompt, workspace, model, skills, and permission. | Host-wide |
| Create from chat | Describe the schedule in any conversation, or select Create in chat. | Current conversation |
| Pause or resume | Use the switch on a task card. | One rule |
| Run now | Open the card menu and select Run now. | One rule |
| Delete | Open the card menu and select Delete task. Run history is kept. | Definition only |
| Inspect runs | Open Run history, then filter by day, week, month, task, or status. | Host-wide |
Each dispatched run uses the saved prompt, workspace, model, and permission boundary. It does not reuse approvals from the source chat.
Safety boundary
| Item | Behavior |
|---|---|
| Permission | Options and the default come directly from the Host permissionPresets service, including custom presets. |
| Tool calls | Follow Host tool availability, permission, and approval checks without an additional fixed tool allowlist or a background-shell ban. The Host manages background processes. |
| Full access | The official danger-full-access option uses the same risk confirmation and orange warning as Chat. |
| Approval | Chat create follows the session policy. Full access (never) proceeds; Workspace Write / Read Only (ask) shows the official card. Unattended runs stay fail-closed never. |
| Retry | No automatic retry after a started run. |
| Host restart | Leftover queued / running records become failed(host_interrupted). |
| Overlap | One active run per rule. A colliding occurrence is recorded as skipped(overlap). |
A schedule stores future intent. It is not a cached permission grant.
Development
Current sources live in src and build into lib:
- src\index.ts: Host plugin, tools, and RPC.
- src\service.ts: Durable definitions, clock, and run admission.
- src\client\index.ts: Settings page and chat prefill.
tests\*.test.ts: Domain, recurrence, service, client, and package-contract tests.
After changing files, run tests, rebuild, and reinstall from the local directory:
[Console]::OutputEncoding = [System.Text.Encoding]::UTF8
$OutputEncoding = [System.Text.Encoding]::UTF8
pnpm check
dsh plugin --profile web add .
Keep the at-most-once dispatch policy, workspace scoping for Agent tools, and fail-closed unattended approval when changing execution code.
Verification
[Console]::OutputEncoding = [System.Text.Encoding]::UTF8
$OutputEncoding = [System.Text.Encoding]::UTF8
pnpm test
pnpm build
pnpm check runs typecheck, tests, and build together.
Project docs and license
Start from the documentation entry for project status, architecture, and the current iteration. Product notes live in NOTICE.
This project uses Apache License 2.0.
Operate deliberately
Install and manage
Prerequisites and target Profile
Target: Web Profile
Delivery: Dsh Bundle Git — MichengAI/dsh-automation#f1bc91a3437f0b952631a46a8363089587b9ae6a。
Verify, update, and remove
Show lifecycle commands
dsh plugin --profile web listCompatibility and access
DeepSeek Harness DSH Web plugin; Node.js 22.19+ for source work: DSH peers: >=0.1.0-rc.5 <0.2.0; Cordis: >=4.0.1 <5.0.0。
Review compatibility evidence ↗
Risk facts
Scheduled runs can execute saved coding prompts in selected workspaces under Host permission presets.
Evidence ↗Each occurrence starts a fresh root Agent and session; source-chat history is not inherited.
Evidence ↗Evidence and editorial reviewManifest, Bundle patch, distribution and freshness
Immutable evidence
Review status and source activity
Use a permission preset appropriate for unattended code execution and review saved prompts, workspace selection, and run history before enabling recurring rules.
AI reviewed Sep 10, 2026, 11:36 AM UTC。GitHub facts last checked Sep 10, 2026, 11:36 AM UTC。
No material source change has been recorded since this evidence baseline.