feat: add docx loremify sanitizer for confidential bug reports

[fahdarafat]

Summary

I hope this helps so issue reporters can share reproducible files without exposing confidential
text.

What this PR changes

• Adds a new CLI script: scripts/loremify-docx.ts
• Adds a package script:
  • bun run docx:loremify -- <file.docx>
• Adds contributor documentation in CONTRIBUTING.md showing how to sanitize files before attaching them to issues

Behavior

• Replaces human-readable DOCX run text with lorem-style text
• Preserves document structure and formatting-critical layout containers (tables, nesting, headers/footers, styles,
  relationships)
• Preserves per-word lengths so visual flow stays close to the source document
• Defaults to safer story-part replacement (main document + headers/footers/notes/comments)
• Supports --all-xml for advanced full-XML replacement when needed

Usage

bun run docx:loremify -- "path/to/file.docx"
bun run docx:loremify -- "path/to/file.docx" --out-dir sanitized-docs

[vercel]

Someone is attempting to deploy a commit to the EigenPal Team on Vercel.

A member of the Team first needs to authorize it.

[jedrazb]

Nice contribution! The code is clean and well-structured. A few suggestions to consider:

1. Document metadata leaks — docProps/core.xml and docProps/app.xml (author, company, manager, revision history) aren't sanitized. This is probably the biggest gap since someone could sanitize text but still ship identifying metadata.

2. Embedded images/media — files in word/media/ survive sanitization. Logos, charts, screenshots can be just as identifying as text. Worth at least documenting this limitation, or adding a --strip-media option.

3. No tests — the core functions (loremifyXml, applyCasing, decodeXmlEntities, parseArgs) are pure and easy to unit test.

4. Numbers become lorem text — the [\p{L}\p{N}]+ regex replaces digits too, so dates like 2026-01-15 become lore-ips-dol. Preserving numeric characters would keep layouts more realistic.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
docx-js-editor	Ready	Preview, Comment	Mar 3, 2026 4:08am