OfficeCLI is a standalone command-line Office suite that lets AI agents read, edit, render, and automate Word, Excel, and PowerPoint files without requiring Microsoft Office. Developers and other users employ it to create and inspect documents through commands, with HTML or PNG rendering supporting visual review. The catalogue skills teach coding agents how to install and use OfficeCLI.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add iOfficeAI/OfficeCLI --skill officecli-word-formgit clone --depth 1 https://github.com/iOfficeAI/OfficeCLIWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/iofficeai/officecli/officecli-word-form)<a href="https://agentmods.dev/skills/iofficeai/officecli/officecli-word-form"><img src="https://agentmods.dev/badge/skills/iofficeai/officecli/officecli-word-form/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/iofficeai/officecli/officecli-word-form"><img src="https://agentmods.dev/badge/skills/iofficeai/officecli/officecli-word-form.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 5 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Supply Chain · line 18 Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.Fix: Avoid downloading and executing remote scripts. Use trusted packages from PyPI/npm. If remote fetch is required, verify checksums and use HTTPS.
- high Tool Misuse · line 18 Tool calls are chained to bypass individual safety checks or escalate capabilities beyond what any single tool call would allow.Fix: Limit tool chaining depth and validate the output of each tool before passing it to the next. Require explicit user approval for multi-step chains.
- high Tool Misuse · line 184 Tool calls are chained to bypass individual safety checks or escalate capabilities beyond what any single tool call would allow.Fix: Limit tool chaining depth and validate the output of each tool before passing it to the next. Require explicit user approval for multi-step chains.
- high Anti-Refusal · line 595 Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.Fix: Remove instructions that suppress warnings, disclaimers, or ethical commentary. Let the agent surface safety-relevant caveats to the user.
- high Anti-Refusal · line 651 Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.Fix: Remove instructions that suppress warnings, disclaimers, or ethical commentary. Let the agent surface safety-relevant caveats to the user.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00224 | $0.12789 |
| Opus 5.5 | $0.00090 | $0.05116 |
| Sonnet 5.5 | $0.00045 | $0.02558 |
| Haiku 4.5 | $0.00022 | $0.01279 |
Grade C, and why
officecli-word-form scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 17d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
curl -fsSL https://d.officecli.ai/install.sh | bash Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -fsSL https://d.officecli.ai/install.sh | bash How it starts
The opening of the file, as written. The whole thing — 677 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OfficeCLI Word-Form Skill
This skill is INDEPENDENT, not a scene layer on docx. A form's payload — <w:sdt> controls, <w:ffData> legacy fields, <w:fldChar> mail-merge, documentProtection — is a distinct element class from docx's paragraph/heading/style primitives. Its QA is different too: docx's Delivery Gate cares about visual layout and live PAGE fields, this skill's cares about data plumbing (protection enforced / alias+tag / items injected / name ≤ 20 / no underscore anti-pattern). Reverse handoff: if the user's document has no fillable fields (report, letter, memo, thesis, proposal), route to officecli-docx or a docx scene skill — don't use this one.
BEFORE YOU START (CRITICAL)
If officecli is not installed:
macOS / Linux
if ! command -v officecli >/dev/null 2>&1; then
curl -fsSL https://d.officecli.ai/install.sh | bash
fi
Windows (PowerShell)
if (-not (Get-Command officecli -ErrorAction SilentlyContinue)) {
irm https://d.officecli.ai/install.ps1 | iex
}
Verify: officecli --version
If officecli is still not found after first install, open a new terminal and run the verify command again.
If the install command above fails (e.g. blocked by security policy, no network access, or insufficient permissions), install manually — download the binary for your platform from https://github.com/iOfficeAI/OfficeCLI/releases — then re-run the verify command.
Help-First Rule
This skill teaches what a real form needs, not every CLI flag. When a prop / alias / enum is uncertain, consult help BEFORE guessing: officecli help docx [element] [--json] (e.g. sdt, formfield, field). Help is pinned to the installed CLI version and is authoritative — when this skill and help disagree, help wins (the prop set on sdt in particular has grown over time; trust help docx sdt, not a hardcoded list).
Mental Model & Inheritance
A Word form is a .docx plus four OpenXML payload layers plain-docx skills do not touch: <w:sdt> content controls (types: text / richtext / dropdown / combobox / date / picture / group), <w:ffData> legacy FormField (still the only way to get a real checkbox — SDT type=checkbox is not implemented), <w:fldChar> complex fields (MERGEFIELD, REF, PAGEREF, SEQ, IF — template-time, not user-fill), and documentProtection (the lock that makes non-field text read-only in Word — and, on the CLI, protection=forms locks non-field content edits (those need --force or raw-set) but still allows form-field (SDT) edits, which is the point of forms protection).
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 17d ago Changed 6c6154954089
- 28d ago Changed c4cc673eccbf
- 1mo ago First seen · 677 lines · 224 tokens per session scan C 428eef196cd5
officecli-word-form is a skill published in the GitHub repository iOfficeAI/OfficeCLI (32,194 stars, last pushed 7d ago), licensed Apache-2.0. It adds 224 tokens to every session and 12,789 once invoked, about $0.0009 per session on Opus 5.5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
consulting-pdf-from-youtube
Generate polished Chinese consulting-style PDF reports from YouTube videos. Uses a transcription priority chain -- YouTube auto-subs first, Whisper fallback second (with quality gates and proper noun verification). Includes bot-detection bypass escalation when yt-dlp is blocked. Use when a user shares a YouTube link…
lark-sheets
A tool for creating and editing Feishu spreadsheets, which are online tables for storing, calculating, and presenting data. It also manages sheets, formulas, formatting, charts, pivot tables, and filters.
chinese-pdf-report
Create professional Chinese PDF reports with reliable font rendering on macOS, especially when previous HTML-to-PDF output produced garbled Chinese text, missing glyphs, or ugly default typography. Use when asked to generate or re-export Chinese-heavy PDFs, fix Chinese font issues, improve typography for formal…
lark-minutes
A tool for finding and working with Feishu meeting recordings and their transcripts or AI-generated notes. Feishu is a workplace collaboration platform, and a meeting transcript is a written record of what people said.
lark-doc
A tool for reading and editing Feishu cloud documents, including document pages and wiki entries. Feishu is a workplace collaboration platform commonly used for shared documents and knowledge bases.
youtube-bilingual-transcript-report
Download a YouTube video, build English and Chinese timestamped transcripts, create a structured summary with independent insights, and export a professional Chinese PDF report. Includes fallback when Chinese subtitle download hits HTTP 429 and when full-resolution video download is unnecessarily heavy.