Anthropic / skills/docx/SKILL.md
DOCX
Creates and edits Word documents through JavaScript generation, XML editing and rendered-output checks.
Runtime probed · reviewed · A helper or bounded runtime behavior was probed. This does not establish model task effectiveness or comprehensive security.
01 / Task fit
Where it may help.
These are intended uses indicated by the upstream material. They are not measured outcomes.
- Planning Word document generation and editing
- Understanding tracked-change and rendering pitfalls
02 / Setup & access
What it asks of your environment.
Requirements
- Creation examples use Node.js and the docx npm package; reading examples use Pandoc.
- Editing and verification use Python, ZIP tools, LibreOffice (soffice) and Poppler's pdftoppm; merge_runs.py imports defusedxml.
- The optional restricted-socket workaround invokes gcc and LD_PRELOAD, a Linux-style path; native Windows compatibility is unverified.
- No dedicated API key is specified for the reviewed local helpers. Hosted-agent access and Anthropic's reuse terms remain separate.
Permissions & data access
- Read and rewrite document XML, extracted directories and output files.
- Launch document conversion processes and create temporary LibreOffice profiles.
- The socket workaround can compile a temporary shared library and preload it into LibreOffice.
Instructions alone do not enforce a sandbox. The agent host, connected tools and approvals determine what the skill can actually read, change or send.
03 / Things to check
Security concerns & practical limits.
These concerns follow from the reviewed sources. They describe possible exposure or required review; they are not confirmed vulnerabilities.
Higher attention
Restricted reuse terms
The document skill has proprietary terms with copying, derivative-work and distribution restrictions. This entry is a source-linked evaluation, not permission to fork or bundle it.
Review before use
Overwrite and native-code behavior
merge_runs.py overwrites a DOCX by default unless given an output path. soffice.py may compile and reuse a shared library in the temp directory. IonBin assessment: preserve originals and review the converter environment before use.
Limits and nuances
- The entrypoint assumes some tooling is preinstalled; this was not confirmed in any target environment.
- The source itself describes tracked-paragraph deletion differences between Pandoc, LibreOffice and Word.
- Validation, archive extraction and all document-editing helpers were not fully reviewed.
04 / Review record
What we checked. What’s still open.
Recorded review scope
- Entire SKILL.md
- scripts/office/soffice.py
- scripts/merge_runs.py imports and document read/write paths
- LICENSE.txt
- Bounded evaluation; exact operations, counts, environment and limitations in the linked public report
Not tested or established
- No effectiveness benchmark or security certification
- Transitive dependencies, current advisories and host-agent permission controls not audited
- No rendered document, Word/LibreOffice interoperability or cross-platform tests
- No complete native-library or XML-validator audit
- Only merge_runs and its safe_extract helper were exercised
- No full DOCX rendering, Word/LibreOffice interoperability, tracked-change acceptance or generated-document quality tests
- Review date
- Source revision
34040c9c568585f6929bedeaad110ad08f079624- Our relationship
- Explored. Investigated through the recorded review. No use or replication is claimed.
Upstream content can change after review. A repository revision identifies source scope; it does not prove that every script, dependency or runtime behavior was tested. How freshness is handled.
Recorded evaluation / Helper probe
Results with their limits attached.
Synthetic merge, formatting/revision boundaries, XML errors and ZIP traversal: 7 of 7 expectations passed. No rendered Word or LibreOffice workflow was tested.
- Passed
- 7
- Failed
- 0
- Total recorded
- 7
These counts refer to the cases in this report. They are not a benchmark against other skills. A helper probe does not measure model task effectiveness or establish comprehensive security.
Read the evidence reportDownload the authored reproduction harness (JSON bundle) ↗
The report provides the case definitions, observations and baseline scope. Missing baselines or untested behaviors remain limitations.
Recorded environment
- Windows-11-10.0.26200-SP0; Python 3.12.10; Node 24.18.0; Chrome 155.0.8059.82
- Synthetic local fixtures; minimal child environment; no live account or credential; no OS sandbox
- Dependencies: pypdf 6.19.0, reportlab 4.4.9, Pillow 12.3.0, pdf2image 1.17.0, defusedxml 0.7.1
- Git Bash 5.3.15(1)-release (x86_64-pc-cygwin); Poppler 26.07.0
Recorded commands (1)
Commands document this evaluation. They are not setup instructions for your environment.
python run_evals.py --runtime-libs <existing-bundled-site-packages> --poppler-bin <poppler-binaries>Findings from this evaluation
No individual findings were recorded. This does not establish the absence of risk.
Evaluation limitations
- Only merge_runs and its safe_extract helper were exercised
- No full DOCX rendering, Word/LibreOffice interoperability, tracked-change acceptance or generated-document quality tests
- Nominal merge probes used unpacked synthetic document XML; traversal used a synthetic ZIP
- No ZIP bombs, arbitrary hostile native code or large-resource tests
- Proprietary/source-available helper code is retained privately and excluded from public artifacts
- Supervised local processes, not an OS sandbox
- No container, VM or memory/resource quota
- Application guards do not establish containment of arbitrary hostile code
- No real credential, remote account, or destructive-effect testing
05 / Reuse & discovery
License, attribution & popularity.
Proprietary / source-available
Use is governed by the per-skill terms and applicable Anthropic agreement. Public visibility does not make this an open-source skill.
Why it is in this collection
Rounded repository-wide GitHub stars observed in the public repository page; not skill installs, a quality score, or a measured trend.
182,000 repository stars (rounded)
- Scope
- Repository-wide; not individual skill usage.
- Observed
- Time window
- Single observation; no trend is established.
Popularity and inclusion do not establish quality, compatibility or security. Counts belong to their named metric and date; they are not a current ranking.
06 / Follow the evidence
Go to the original.
- Document tasks, tools and tracked-change caveats github.com
- LibreOffice process, temporary profile and preload behavior github.com
- Python dependency and default overwrite behavior github.com
- Per-skill reuse restrictions github.com
- Source-available distinction and repository popularity github.com
IonBin’s source notes are independent assessments. Upstream authors retain their work and attribution. This page does not install or execute the skill.