A macOS-specific skill for AI agents to control live Microsoft Word applications, handling selections, comments, and review states via osascript.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install word
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install word using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Word?
The Word skill is designed for AI agents that need to interact with Microsoft Word as a live, functional application rather than a static file format. By utilizing Openclaw Skills, developers can enable agents to attach to active sessions, manage document states, and perform selection-aware edits directly within the Word interface on macOS. This approach ensures that track changes, comments, and complex document metadata are preserved and handled accurately through the official osascript CLI.
This skill bridges the gap between raw document manipulation and live application control, making it ideal for workflows that require human-in-the-loop review or real-time document interaction. Whether it is managing a multi-document review cycle or automating exports with updated fields, this skill provides the necessary patterns for reliable Word automation.
Word Use Cases
- Automating document review by adding, reading, or accepting tracked changes and comments.
- Performing live document edits that require awareness of the current cursor position or selection range.
- Batch exporting documents to PDF or other formats while ensuring tables of contents and fields are updated.
- Standardizing document cleanup by removing comments or resolving tracked changes before final delivery.
- Integrating Word automation into complex macOS workflows using Openclaw Skills to maintain session context.
How Word Works
- The AI agent identifies whether the target document is already open or needs to be launched using the osascript interface.
- The skill performs a context check to verify the active document name, current selection range, and track-change status.
- Commands are sent to the Microsoft Word scripting dictionary to execute specific actions such as text insertion, comment creation, or view adjustments.
- For high-risk actions like accepting all changes, the skill follows a safety protocol to request explicit confirmation.
- Post-execution verification is performed by re-reading the document state to ensure the requested changes were applied correctly.
- The skill manages the lifecycle of the Word session, ensuring that only agent-created documents are closed upon completion, leaving user-owned documents untouched.
Word Setup
To get started with this skill within the Openclaw Skills ecosystem, ensure you are running macOS with Microsoft Word installed. Run the following commands to initialize the local memory and environment:
mkdir -p ~/word/
# Install via the clawhub registry
clawhub install word
After installation, the agent will reference the configuration paths and safe defaults established in your local directory.
Word Data Schema & Taxonomy
The skill organizes its persistent state and configuration within the ~/word/ directory. This data architecture ensures that Openclaw Skills maintain context across different agent sessions.
| File | Purpose |
|---|---|
memory.md |
Contains environment facts, safe defaults, and the last working control path used by the agent. |
incidents.md |
Logs reusable failures and proven recovery steps to improve future automation reliability. |
document-notes.md |
Stores non-sensitive notes about trusted documents, preferred views, and target export paths. |
Word Advanced Features
- Selection-aware safety logic that anchors actions to specific document objects rather than just the insertion point.
- Automated recovery patterns that detect and report blockers like modal dialogs, protected views, or missing permissions.
- Support for complex story ranges, allowing the agent to interact with headers, footers, and text boxes.
- Detailed provenance reporting to track whether documents were pre-existing or opened specifically for a task.
- Native integration with other Openclaw Skills for cross-application Office automation and advanced scripting workflows.
SKILL.md
Loading
Related Openclaw Skills
Maps
A sophisticated tool for cross-provider map workflows including geocoding, routing, and spatial data normalization.
Product Image Upscaler
Enhance the clarity and resolution of ecommerce product images to make them commercially viable for listings, PDPs, and digital ads.
Audio to Text Caption
A high-efficiency transcription and captioning workflow designed for creators and e-commerce teams.
TikTok Affiliate ROI Calculator
A high-performance financial modeling tool designed to estimate TikTok affiliate campaign profitability and risk before launch.
Forguncy Web Operator
A specialized tool for programmatically interacting with and extracting data from SpreadJS tables within Forguncy web applications.
AI Cinema Director
A specialized tool that translates high-level creative intent into professional technical directives for state-of-the-art AI video models.
Featured*
