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.

Offpeak Job — DSH Plugin for DeepSeek Harness
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
O

dsh-offpeak-job

Offpeak Job

Session off-peak toggle + multiple provider-configured off-peak windows

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

npx -y @deepseek-ai/dsh plugin --profile web add github:ganningniang/dsh-offpeak-job#621e5035f9bc486df461b6177ff29a3bd6fcd8e4
READMECompatibilityVersions
Peak / off-peak settings

Compatibility and provenance

Offpeak Job is published as dsh-offpeak-job and currently resolves to version 0.3.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
web
Release source
github
Registry updated
9/17/2026

Versions

0.3.0stable
9/17/2026

Related plugins

Loading related plugins…

Latest
0.3.0
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
web
License
MIT
Source
github
GitHub
★ 0
Weekly downloads
0
Last push
9/17/2026
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

Related plugins

More verified plugins in models-usage.

Usage@linxin666/dsh-usageUsage statistics plugin for the dsh web GUI: per-provider balance and coding-plan quota detection plus a live token usage ledger, with the current session provider's today usage on the sidebar entryUsage Stats@ychris12138/dsh-usage-statsToken usage heatmap, provider balances, and subscription quotas for the dsh web GUICodex Connectdsh-codex-connectChatGPT OAuth and Codex models for DeepSeek Harness.Ui Usage Billing@kenz1117/dsh-ui-usage-billingUsage billing dashboard for DeepSeek Harness: sidebar cost metrics plus a full dashboard modal, priced from a current multi-provider catalog with real usage aggregated from session logs.

README

dsh-offpeak-job 🌙

English · 简体中文

Off-peak scheduling for DeepSeek Harness — park chats until cheaper hours, pause at peak if you want, and keep an eye on waiting work from the sidebar.

📸 Screenshots

Peak / off-peak settings⚙️ Peak / off-peak settings — per provider: custom windows, what to do when peak starts, and how escalations behave on overnight runs

✨ Feature tour

⚙️ Peak / off-peak per provider

  • Lives under Settings → Models on each provider card
  • DeepSeek can use the official peak calendar; or define your own windows (every day / weekdays / weekends), including overnight spans like 18:00–09:00
  • Turn peak / off-peak off entirely when you do not want waiting at all
  • Optional: pause when peak begins, resume next off-peak — or keep a started turn running through peak
  • Optional: overnight escalation policy — reject and continue / wait for you / auto-allow (Off-peak run only)

💬 Run timing per chat

  • Composer control: Off-peak run / Run now (same place as access mode; hideable from provider settings)
  • Outside the window, messages wait instead of spinning the chat the whole time
  • Start now on the wait bar runs the queued turn immediately without flipping the chat back to Run now
  • Waiting messages and “continue later” work survive Host restart

📋 Off-peak overview

  • Sidebar entry above Settings (hidden when every provider has peak / off-peak off)
  • Cross-chat board of Off-peak-run sessions: waiting, paused for later, needs you, session unavailable
  • Start now / discard from one place when something is waiting for a window

At a glance

🧩 Plugin typeCordis plugin — host routes + web client, pure additive (no official plugin replaced)
🌙 What it gatesAdmission of sends on Off-peak-run chats; optional pause at window end
💬 Chat UXWait bar in the conversation; run-timing control in the composer
📡 Status boardSidebar Off-peak overview with badge for items that need attention
💾 StateUnder $DSH_HOME/offpeak-job/ (provider schedules, per-chat flags, waiting queues)
🎨 UITypeScript + React (host externals) + CSS, follows host chrome

Quick start

  1. Install (pick one):

    A · one-liner (recommended) — straight from the GitHub Release tarball, no Git needed:

    dsh plugin --profile web add "https://github.com/ganningniang/dsh-offpeak-job/releases/latest/download/dsh-offpeak-job.tgz"
    

    B · local clone (for hacking on the source) — link: accepts a local absolute path only (no spaces in the path):

    git clone https://github.com/ganningniang/dsh-offpeak-job.git
    cd dsh-offpeak-job
    npm install && npm run build
    dsh plugin --profile web add "link:<absolute path of the cloned dsh-offpeak-job directory>"
    # e.g. cloned into D:\tools → dsh plugin --profile web add "link:D:/tools/dsh-offpeak-job"
    

    Either way the add command registers dsh-offpeak-job in the profile bundle list (writes to ~/.dsh, may ask for authorization). If the dsh command is missing, use npx @deepseek-ai/dsh instead.

  2. Restart the DSH web process, refresh the GUI

  3. Configure a provider: Settings → Models → edit a provider → Peak / off-peak → pick windows / pause policy / approval policy → Save

  4. Use it in a chat: choose Off-peak run in the composer; send as usual — work waits until the next open window (or use Start now when you cannot wait)


Architecture

One package ships the host Cordis plugin and the web client:

  • host: /api/offpeak-job/* routes for health, provider/session config, overview, release / discard / send-conflict resolve; wraps agent send paths so Off-peak-run work can wait without opening a busy turn
  • client: injects the Models peak / off-peak panel, composer run-timing control, conversation wait bar, and sidebar overview
  • persistence: schedules and waiting queues live under $DSH_HOME/offpeak-job/; Host restart rehydrates queues and re-arms release timers when the chat agent is live again

Development & testing

cd dsh-offpeak-job
npm install
npm run build     # lib/index.js + lib/client.js
npm run check
npm run test:offpeak
  • Build in this package directory — output is lib/ next to cordis.patch.yml
  • Tests cover schedules, wait persistence, approval policy, overview, and locale freeze

Troubleshooting

Q: How do I reinstall / get the latest release?

Re-run the install command (always installs the latest Release), then restart dsh web and refresh:

dsh plugin --profile web add "https://github.com/ganningniang/dsh-offpeak-job/releases/latest/download/dsh-offpeak-job.tgz"

Queues and peak / off-peak config live under $DSH_HOME/offpeak-job/; reinstalling the plugin usually leaves them alone.

Q: The run-timing control is gray / missing?

The provider has peak / off-peak turned off (or set to “no restriction”), or Show run timing in chat is off in that provider’s peak / off-peak panel. Turn the schedule on and save; enable the composer control if you want it visible.

Q: Off-peak time arrived, but nothing ran?

DeepSeek Harness (and the machine) must still be awake. Sleep, lid close, or quitting the web process means queued work will not start by itself. After restart, waiting items should still show in the wait bar / overview once the chat is available again.

Q: I paused at peak — will it pick up exactly where a long tool left off?

No. Resume means “continue this task” in the next off-peak window, not a full mid-tool freeze. If you hit Stop yourself, nothing auto-resumes.


Known limits

  • The Host must be up during off-peak for queued work to actually run
  • Pause-and-continue is a follow-up to keep going — not an exact checkpoint of mid-tool work
  • Default overnight escalation is reject and continue (change per provider if you want approvals or auto-allow)
  • Only Off-peak run chats follow the calendar and the off-peak approval policy; Run now uses the normal flow
  • Sending a new message while a resume is waiting asks you to finish the resume first or drop it
  • Standalone plugin: does not patch DeepSeek Harness core; if a host extension point is missing, the plugin works around it locally or documents the gap

Privacy

No telemetry. Config and waiting queues stay on disk under $DSH_HOME/offpeak-job/. Network use is only what DeepSeek Harness already does for model calls — this plugin does not phone home.


License

MIT