AI Agent Hub
Back to plugins
🖥️

dsh-yzj

Client Updated 2026.09.03

Run the following command in DeepSeek Harness:

dsh plugin install GuoxinShan/dsh-yzj

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

Run dsh plugin install GuoxinShan/dsh-yzj in DeepSeek Harness to install; the source code is available at https://github.com/GuoxinShan/dsh-yzj . After installation, restart the GUI to see the Yunzhijia entry in the sidebar.

About this plugin

If you manage documents, calendars, spreadsheets, and team messages in Yunzhijia every day and want an AI assistant to work alongside those workflows, dsh-yzj is the bridge. It exposes the full yzj-cli surface as 53 model-facing tools inside DeepSeek Harness—block-level document editing, multi-dimensional table CRUD, calendar participant lookups, IM message recall—so you never have to leave the chat interface.

Security is treated as a first-class concern. The bridge spawns yzj-cli through a bounded subprocess channel; harness never touches appSecret or accessToken. All 31 write tools are gated behind tiered confirmation cards—destructive actions trigger a strong-confirmation red card with full parameter display and no truncation—so you can confirm, cancel, inspect context, or edit before anything executes. The browser workspace folds conversations, calendars, and the knowledge base into a single three-tab panel, and tool results render as rich cards instead of raw CLI logs.

It fits teams that run collaboration on Yunzhijia and want to weave AI into that daily loop: let the assistant look up documents, create calendar entries, or send group messages, with every write operation still requiring your explicit sign-off rather than firing silently.

Use Cases

  • Let the AI assistant manage Yunzhijia documents, calendars, tables, and team messages directly from chat
  • Every write operation passes through tiered confirmation cards—destructive actions trigger strong-confirmation red cards—preventing silent AI execution
  • Review rich tool-result cards and domain-specific confirmation cards in a browser panel with confirm, cancel, inspect context, and edit actions

Best For

  • Development and product teams that run collaboration on Yunzhijia
  • Teams wanting AI in their daily workflow without sacrificing operational safety
  • Enterprise users who need explicit confirmation gates on every AI write action to prevent silent mutations