AI Agent Hub
Back to plugins
🖥️

dsh-doc-quick

Client Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install QinpanWan/dsh-doc-quick

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

Run dsh plugin install QinpanWan/dsh-doc-quick in the DeepSeek Harness terminal to install this plugin; the full source is available at https://github.com/QinpanWan/dsh-doc-quick

About this plugin

Handling local documents inside the dsh web chat has always been hampered by sandbox isolation—the file sits right on your disk, yet the agent cannot reach it, forcing you to paste content manually or route everything through an upload step. dsh-doc-quick closes that gap: drag a file into the dialog and the host process reads it directly from your local filesystem, bypassing the sandbox entirely, so the agent gets full context with zero transfer overhead.

The plugin is written in pure JavaScript with no native build dependencies. It supports PDF, Word (doc/docx), PowerPoint (ppt/pptx), EPUB, Jupyter Notebook, Markdown, and plain text; PDF parsing lazily loads pdfjs-dist when available and degrades gracefully when it is not. On the agent side, three tools—doc_list, doc_read, and doc_complete—let it inspect pending tasks, read file content, and register finished outputs. Once processing is done, a sidebar appears on the right side of the chat, listing task status, summaries, and file paths with preview and download links.

If your daily workflow involves handing office documents to an agent for summarization, rewriting, format conversion, or knowledge extraction, dsh-doc-quick compresses all of that into a single drag-and-drop gesture, eliminating the tedious back-and-forth of moving files between environments.

Use Cases

  • Drop a PDF report into the chat and have the agent extract key data points into a summary
  • Drag a Word contract into the dialog and ask the agent to outline its clause highlights
  • Drop a Jupyter Notebook and have the agent interpret the analysis results and distill conclusions

Best For

  • dsh web users who regularly rely on agents to process office documents
  • Developers frustrated by sandbox isolation who want agents to read local files directly
  • Knowledge workers who prefer a drag-and-drop workflow with minimal file-shuffling overhead