AI Agent Hub
Back to plugins
🖥️

dsh-notifier

Client Updated 2026.09.02

Run the following command in DeepSeek Harness:

dsh plugin install a23842/dsh-notifier

Paste the following prompt into your AI chat to install this plugin:

Install the plugin in DeepSeek Harness by running dsh plugin install a23842/dsh-notifier; the source repository is at https://github.com/a23842/dsh-notifier

About this plugin

A DeepSeek Harness run can take several minutes, and it may hit blocking events along the way-questions to answer, approvals to grant. Switch to another tab or walk away, and those events slip by silently. dsh-notifier exists to close that gap: it makes the Harness reach out to you the moment something needs your attention.

The plugin works on three levels. First, external channel delivery covers NotifyX, WeChat Work app notifications, WeChat Work bots, Resend email, SMTP personal mail, DingTalk bots, and Feishu bots-seven channels pushed directly by the Host process with no dependency on a browser page. Second, browser local notifications fire via the operating-system notification API when the tab is in the background or unfocused; they stay quiet while the page is visible and focused, and sub-agent sessions are excluded. Third, a send_notification model tool lets the running model compose a custom title and body and broadcast it to every enabled channel mid-task. A dedicated Notifications page in DSH Web settings provides per-channel toggles, granular switches for run completion, questions, approvals, and goal completion, and stores all secret fields server-side so they are never echoed back in settings API responses.

It is built for individual developers and teams who run DeepSeek Harness in production or daily work and want to stay in the loop through the chat tools and inboxes they already use-WeChat Work, DingTalk, Feishu, or plain email-rather than keeping a terminal window pinned to their screen. Teams that want the model to proactively notify collaborators mid-task will find the send_notification tool a natural fit as well.

Use Cases

  • Receive automatic pushes on WeChat Work, DingTalk, Feishu, or email when a run finishes or hits a question or approval
  • Get browser OS notifications when the tab is in the background and a run completes, a question appears, or an approval is pending
  • Let the model call send_notification mid-task to broadcast a custom title and body to every enabled channel

Best For

  • Solo developers who run long Harness tasks and cannot keep a terminal window in front of them at all times
  • Ops and collaboration teams that prefer critical events to arrive through WeChat Work, DingTalk, Feishu, or email
  • Workflow builders who want the model to proactively notify collaborators mid-task instead of relying on manual polling