The latest articles on DEV Community by breath57 (@breath57). https://dev.to/breath57 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3809862%2F1722541a-68b6-4940-8dd2-f40af78e4b4b.jpeg https://dev.to/breath57 en breath57 Fri, 06 Mar 2026 12:30:32 +0000 https://dev.to/breath57/i-built-an-open-source-skill-pack-that-lets-ai-agents-operate-dingtalk-for-you-jcp https://dev.to/breath57/i-built-an-open-source-skill-pack-that-lets-ai-agents-operate-dingtalk-for-you-jcp <p>If your company uses DingTalk, you know the drill: meeting wraps up, now you have to manually archive the notes, update the project table, log the action items...</p> <p>It's not hard. It's just <em>exhausting</em>.</p> <p><strong>What if your AI assistant could do it all?</strong></p> <h2> What this project does </h2> <p><a href="https://github.com/breath57/dingtalk-skills" rel="noopener noreferrer"><strong>dingtalk-skills</strong></a> is an open-source skill pack that gives Claude Code, GitHub Copilot, Cursor, and other AI agents the ability to directly operate DingTalk — documents, knowledge bases, and AI tables.</p> <p>After installing, you can just tell your AI:</p> <blockquote> <p>"Save today's meeting notes to the '2026-03' folder in the 'Project Docs' knowledge base"</p> </blockquote> <p>The agent will find the folder, create the doc, write the content. Done.</p> <blockquote> <p>"Show me all records in the requirements table where priority is 'High'"</p> </blockquote> <p>It scrolls through everything and gives you the filtered results.</p> <blockquote> <p>"Add three tasks to the task table — I'll tell you the title, owner, and deadline for each"</p> </blockquote> <p>One sentence per task. Done.</p> <h2> Why it's useful </h2> <p><strong>No API knowledge needed.</strong> Install it once, then just talk naturally. The agent handles the rest.</p> <p><strong>One-time setup.</strong> The first time you use it, the agent walks you through your app credentials and saves them. After that, it works automatically every session.</p> <p><strong>Zero extra dependencies.</strong> No libraries to install. No new packages in your project. It just works.</p> <h2> What's supported </h2> <p><strong>DingTalk Docs &amp; Knowledge Base</strong></p> <ul> <li>List your knowledge bases</li> <li>Browse directory structure</li> <li>Create documents and folders</li> <li>Read, write, and update document content</li> <li>Delete documents</li> <li>Manage member permissions</li> </ul> <p><strong>DingTalk AI Table</strong></p> <ul> <li>View, create, and delete worksheets</li> <li>Manage fields</li> <li>Batch insert records</li> <li>Query all records (with pagination)</li> <li>Update and delete specific records</li> </ul> <h2> How to install </h2> <p>Prerequisite: a DingTalk internal app on the open platform with document/table permissions enabled.<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>npx skills add breath57/dingtalk-skills@dingtalk-document npx skills add breath57/dingtalk-skills@dingtalk-ai-table </code></pre> </div> <p>That's it. Talk to your agent and it'll guide you through first-time setup automatically.</p> <p>More scenarios are planned: messages, approvals, contacts, calendar, to-dos.</p> <p>GitHub: <a href="https://github.com/breath57/dingtalk-skills" rel="noopener noreferrer">github.com/breath57/dingtalk-skills</a></p> <p>Stars and issues welcome!</p> ai opensource agents productivity