DOCX Document Skill
Guides Claude through creating, reading, editing, and validating Word documents.
The skill scopes itself to DOCX/DOTX and includes symlink rejection and archive path-escape checks, while operating mainly on local files. However, editing and merge commands default to overwriting inputs, with no user-confirmation, backup, or rollback mechanism; dependency and data-flow disclosure are incomplete, so points are deducted.
The instructions cover creation, reading, editing, comments, and tracked changes, and the scripts include several input checks and error messages. However, there is no committed test suite or CI evidence, dependencies are not version-pinned, and LibreOffice macros, a fixed temporary profile, and abnormal paths are unverified; the static score therefore stays below 10.
Trigger conditions, DOCX/DOTX scope, and exclusions for PDFs, spreadsheets, and unrelated coding tasks are relatively clear, with multiple workflows described. Non-fit boundaries, Chinese-language behavior, and mainland-China environment fit are not documented, and some capabilities depend on a local toolchain, so points are deducted.
The SKILL.md has clear sections for creation, editing, comments, and dependencies, with useful gotchas, commands, and some limitations. However, the license only points to an unprovided LICENSE.txt, and versioning, changelog, maintenance ownership, update path, systematic examples, and FAQs are absent, so points are deducted.
The skill provides directly usable commands, script entry points, and DOCX-specific guidance that could support common creation, conversion, editing, and commenting tasks. Static evidence contains no verified representative outputs or test results; some results require manual inspection and tool availability is uncertain, so the score is conservatively limited.
The source code, command-line interfaces, input checks, and some safety logic are auditable and support limited static review. There is no test suite, CI, execution log, or third-party reproduction evidence, so key paths and cross-platform behavior cannot be independently verified; the score remains below 5.
- Editing and merge operations may overwrite input files by default; create a backup and confirm output paths before use.
- Versions of docx, pandoc, LibreOffice, and pdftoppm are not pinned or shown to be available in the target environment.
- The comment script creates comment definitions but requires manual marker insertion into document.xml; comments may remain invisible without those markers.
- Complete license, version, and maintenance information is not present in the supplied material.
What does this skill do, and when should you use it?
This Word-document skill is part of anthropics/skills and targets .docx and .dotx files. It covers new document creation, existing-document editing, content extraction, tracked changes, comments, and image replacement. The instructions rely on local scripts, command-line tools, and filesystem access, while documenting common implementation pitfalls. The repository presents these document skills as source-available reference material rather than open source.
For creation, it uses the preinstalled docx npm package through a script; for editing, it unzips a DOCX, modifies word/document.xml, and rezips it; for reading, it converts the file to Markdown with pandoc. It uses LibreOffice for conversion and validation and pdftoppm for rendering PDF pages to images for inspection. Repository scripts can merge runs, accept tracked changes, and add comments.
- A user needs a polished Word report with headings, a table of contents, page numbers, tables, or a letterhead.
- A user needs to extract and reorganize content from an existing .docx file.
- A user needs to perform find-and-replace operations in Word XML.
- A user needs to manage tracked changes, accept revisions, or add document comments.
- A user needs to insert or replace images in a Word document.
What are this skill's strengths and limitations?
- Covers creation, reading, editing, conversion, validation, tracked changes, and comments.
- Documents concrete implementation pitfalls for tables, page sizes, lists, tables of contents, and images.
- Includes scripts for merging runs, accepting changes, and adding comments.
- Provides a post-generation conversion, rendering, and inspection workflow.
- The SKILL.md lists a Proprietary license; the README describes the document skills as source-available rather than open source.
- It depends on local tools including Node.js, Pandoc, LibreOffice, pdftoppm, and Python.
- The source provides no test-suite, platform-coverage, or performance evidence.
- Editing existing documents requires direct manipulation of Word XML, which is comparatively complex.
How do you install this skill?
The README documents Claude Code installation: run /plugin marketplace add anthropics/skills, then install the document-skills plugin. Alternatively, run /plugin install document-skills@anthropic-agent-skills. The source does not document a separate installation procedure for other clients.
How do you use this skill?
After installation, provide a Word-document request such as: Create a polished Word report with headings, a table of contents, page numbers, and a letterhead. You can also mention .docx, .dotx, tracked changes, comments, or image replacement explicitly.