docx-editor-agents has a great API for AI document review. Thank you!
I would like to use this library as part of an agent skill for Claude Code (or Cowork within Claude Desktop).
I know there's an MCP, but:
- it seems the MCP doesn't expose all the tools, and
- I'm starting to prefer CLIs for use by agents
Would you consider a PR for a thin CLI wrapper that exposes the existing DocxReviewer API as shell commands:
docx-agent read contract.docxdocx-agent comment contract.docx 5 "Liability cap seems too low"docx-agent replace contract.docx 5 '$50k' '$500k'- ...
Each command would be stateless: read file, create DocxReviewer, call one method, print output, write file back if needed. Plain text output by default, --json flag for structured output.
I am thinking about:
- Creating a PR for this, OR
- Creating a separate package that depends on @eigenpal/docx-editor-agents, OR
- Being convinced that Claude Code should just write a short JS script for each operation
docx-editor-agents has a great API for AI document review. Thank you!
I would like to use this library as part of an agent skill for Claude Code (or Cowork within Claude Desktop).
I know there's an MCP, but:
Would you consider a PR for a thin CLI wrapper that exposes the existing DocxReviewer API as shell commands:
docx-agent read contract.docxdocx-agent comment contract.docx 5 "Liability cap seems too low"docx-agent replace contract.docx 5 '$50k' '$500k'Each command would be stateless: read file, create DocxReviewer, call one method, print output, write file back if needed. Plain text output by default, --json flag for structured output.
I am thinking about: