OfficeCLI DOCX Skill
Give AI agents full control over Word documents via a command-line interface — no Microsoft Office required.
The skill instructs the user to install a binary via curl/irm pipes, which carries supply-chain risk, but provides clear rollback and verification methods (e.g., officecli --version), emphasizes data-flow transparency (open/save lifecycle, flushing before non-officecli reads), and shows no obvious permission overreach or exfiltration. However, since the publisher is unverified and the install executes remote code, trust is deducted. Evidence shows main risks visible but permissions, confirmation, isolation, or recovery incomplete, hence 15 points.
The skill provides clear command examples, error handling advice (step-by-step execution, checking exit codes), and QA processes (view issues, validate), but lacks executable test suites or CI verification results; key-path reproducibility cannot be confirmed from documentation alone. Static review cannot verify actual execution, so reliability scores 10.
The skill clearly defines trigger conditions (when .docx is involved), use cases (create, read, edit Word docs), capability boundaries (e.g., cannot compute TOC page numbers, requires Word refresh), and environment fit (multi-language support including Chinese, but install scripts depend on overseas servers d.officecli.ai and github.com, which may be unreachable from mainland China). Though scenarios are clear, boundary and environment evidence are limited, so 12 points.
Documentation is well-structured with installation, parameters, examples, and FAQs, and has version/release management via CI and npm, but lacks clear maintenance responsibility and update path, and publisher identity is unverified. Readable but incomplete governance, hence 10.
The skill claims to complete tasks like creating, reading, editing Word docs, with examples and verification steps, but static review cannot verify direct usability of outputs, and comparative benefit over manual/alternatives is not evidenced. Per static rule, effectiveness caps at 7.
Documentation provides some auditable primary material (e.g., CI workflows with smoke tests covering create/add/get/close but not critical skill paths like TOC, fields), but no committed test suite for reproduction. Hence verifiability caps at 5.
- The install script executes remote code via curl/irm pipes, posing supply-chain risk; verify publisher identity or manually verify hashes after download.
- Core functionality depends on overseas servers (d.officecli.ai and github.com), which may be unreachable from mainland China, affecting installation and updates.
- Examples and validation claims in the documentation have not been independently tested; verify before use.
- The publisher is unverified; identity is unknown, so proceed with caution.
What does this skill do, and when should you use it?
The OfficeCLI DOCX skill equips AI agents with a command-line tool to create, read, and edit Word documents through semantic paths and properties rather than raw XML. It supports creating documents from scratch, extracting text and structure, modifying formatting, and adding tables, images, headers/footers, TOCs, fields, and more. With a built-in rendering engine and QA tools, it ensures output documents meet professional visual and structural standards. The skill emphasizes help-first discovery, incremental execution, and strict quality gates, making it suitable for agents automating document generation or modification.
This skill drives the OfficeCLI command-line tool to: create new .docx files; read document structure (outline), text, properties, and statistics; edit paragraphs, styles, tables, images, charts, fields, headers/footers, and TOCs; support batch operations, raw XML manipulation, and validation. It operates via semantic paths (e.g., /body/p[1]) and properties (e.g., text, style, size), and provides extensive help commands and query capabilities.
- A data analyst needs to extract text and structure from multiple .docx reports for content analysis or summarization.
- A content creator wants to automatically generate professionally formatted reports or proposals from templates.
- A developer in a CI/CD pipeline batch-modifies or generates Word documents using scripts.
- An AI agent generates a document and wants to visually verify layout and auto-fix formatting issues via screenshots.
- An HR team processes resumes or contract templates, replacing placeholders and generating final files.
What are this skill's strengths and limitations?
- No Microsoft Office installation needed, single binary, cross-platform.
- Semantic command-line interface simpler than raw XML manipulation.
- Built-in rendering engine for HTML/PNG output, enabling AI visual verification.
- Robust QA tools including validation, issue lists, and visual checks.
- Full support for complex document structures: tables, charts, fields, TOCs, etc.
- Requires an additional CLI dependency.
- Complex formatting (e.g., advanced math) may require raw XML, with a steeper learning curve.
- Some features (like TOC page numbers) rely on a Word-compatible field engine, otherwise need manual intervention.
- No test suite or coverage of Office versions is documented.
How do you install this skill?
The skill requires the OfficeCLI binary. Install via: macOS/Linux curl -fsSL https://d.officecli.ai/install.sh | bash, Windows PowerShell irm https://d.officecli.ai/install.ps1 | iex. If that fails, download a binary from GitHub Releases. The skill file resides in the repository at skills/officecli-docx/SKILL.md; provide its content to your AI agent.
How do you use this skill?
Feed the skill description and SKILL.md content to your agent. Basic workflow: set a file variable FILE="your.docx", create with officecli create, open with officecli open, add content with officecli add, modify properties with officecli set, then save with officecli save and validate with officecli validate. Example: officecli add "$FILE" /body --type paragraph --prop text="Hello" --prop style=Heading1.