AI Agent Hub
Back to plugins
🧰

dsh-notion-skill

Web Tools Updated 2026.08.27

Run the following command in DeepSeek Harness:

dsh plugin install Zhiyi-Zhao/dsh-notion-skill

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

Run dsh plugin install Zhiyi-Zhao/dsh-notion-skill in your DeepSeek Harness terminal to install; source code available at https://github.com/Zhiyi-Zhao/dsh-notion-skill

About this plugin

If you rely on Notion daily for notes, tasks, and project databases but keep switching between the app and your agent to search, copy, and update entries, dsh-notion-skill eliminates that friction. It lets a DeepSeek Harness agent talk to your Notion workspace through the official REST API, so a single natural-language prompt can search, read, create, and update content without manual window-hopping.

The skill covers a full read-write loop: search pages and databases; recursively read page content and convert it to Markdown (headings, lists, code blocks, quotes, and more); create sub-pages, append content blocks, and update titles; query, create databases, and add or update database entries. Every write action requires the agent to present a summary and receive user confirmation before executing. The authentication token lives only in a local config file and is never committed to code. Everything is built on the pure Python standard library with zero third-party dependencies, running the same way on Windows, macOS, and Linux.

It is well suited for heavy Notion users who want natural-language control over their workspace, and for developers who need a lightweight, dependency-free bridge between an agent and Notion without installing additional SDKs.

Use Cases

  • Search and read Notion pages or databases using natural language
  • Let the agent write conversation output into a new Notion page
  • Batch query and update project entries in a Notion database

Best For

  • Heavy Notion users wanting to reduce manual window-switching
  • Developers driving workflows with DSH agents
  • Those needing light integration without extra SDKs