dsh-office-tools
Run the following command in DeepSeek Harness:
dsh plugin install kw78/dsh-office-tools
Paste the following prompt into your AI chat to install this plugin:
Please install the plugin and restart the DeepSeek Harness service using the complete source address https://github.com/kw78/dsh-office-tools, after which the agent workflows will gain immediate access to office document parsing and generation capabilities.
About this plugin
Handling office documents within AI-driven workflows often poses significant performance and security challenges. Traditional approaches frequently rely on heavyweight external processes or complex dependency chains, which can lead to environment conflicts and path privilege vulnerabilities. dsh-office-tools is purpose-built for the DeepSeek Harness ecosystem, embedding full Word, Excel, and PowerPoint processing capabilities directly into the plugin host. By leveraging pure JavaScript libraries instead of spawning system-level software subprocesses, it eliminates sandbox escape risks while delivering robust document manipulation out of the box.
The plugin offers seven lightweight, model-facing tools covering creation, parsing, and updates for .docx, .xlsx, and .pptx files. It supports extracting plain text, generating structured documents with tables and lists, managing multi-sheet Excel grids with precise A1-cell targeting, and building presentation slides complete with titles, notes, and embedded media. Every file operation is strictly confined to the agent’s session workspace, featuring safe relative path resolution and symlink interception. Reads are capped at 50 MiB, and writes employ atomic temp-file renaming with overwrite protection by default, ensuring predictable memory usage and data integrity.
This extension is ideal for developers and AI agent teams looking to integrate seamless office automation into their DeepSeek Harness projects. Adhering to the standard DSH plugin contract, it provides clearly defined parameters, validated output schemas, and clean text projections. By handing models the ability to read and generate documents safely within a resource-bounded environment, it serves as a reliable foundation for building enterprise-grade report generation, data extraction, and automated documentation pipelines.
Use Cases
- Automatically extract key metrics from historical business reports and generate structured Excel tables.
- Batch-format standard PowerPoint presentations featuring titles, paragraphs, and speaker notes for internal reviews.
- Safely read and write local contracts and operational data files within a strictly isolated session workspace.
Best For
- Software developers aiming to seamlessly extend LLM capabilities into enterprise document processing pipelines.
- Engineering teams prioritizing zero external subprocess dependencies and lightweight deployment architectures.
- AI architects focused on building highly secure sandbox environments and controllable reasoning workflows.
Related Plugins
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.