Openclaw Officecli logo

Openclaw Officecli

Organization
officecli
openclaw-officecli

Use when an OpenClaw user clearly wants a local Office or image artifact such as a PPTX, DOCX, XLSX, Report, or IMG, and route the request through officecli agent-bridge instead of parsing human CLI output.

Overview

Publisherofficecli
Repositoryofficecli
Skill nameopenclaw-officecli
Stars
112
Forks
8
Bundled files
7
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 7 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by officecli on GitHub. Read the source before you install it.

Installation

Install the Openclaw Officecli AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/officecli/officecli/tree/main/plugins/openclaw-officecli/skills/openclaw-officecli
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/officecli/officecli.git /tmp/officecli
mkdir -p .claude/skills
cp -r /tmp/officecli/plugins/openclaw-officecli/skills/openclaw-officecli .claude/skills/openclaw-officecli
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Openclaw Officecli in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Openclaw Officecli on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Openclaw Officecli is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

OpenClaw OfficeCLI Skill

Purpose

This skill lets an OpenClaw agent generate local pptx, docx, xlsx, report, and standalone img files through officecli agent-bridge, then send the generated file back to the current channel as an attachment.

Trigger Rules

Trigger when the user clearly wants a file artifact, for example:

  • generate a five-slide PPT about an enterprise collaboration platform
  • write a customer-facing docx for me
  • create a budget excel sheet
  • generate a launch image
  • turn this into slides
  • write a docx for customers

Do not trigger for:

  • pure explanation
  • brainstorming
  • outline-only requests
  • analysis without a file deliverable

Runtime Contract

The skill must use officecli agent-bridge as the local execution protocol.

Do not treat officecli new ... stdout as a protocol.

Always prefer the structured bridge:

  • transport: stdio
  • framing: Content-Length
  • protocol: JSON-RPC 2.0
  • tool: office.generate

Required bridge methods:

  • initialize
  • capabilities/get
  • session/open
  • task/invoke
  • task/respond
  • task/status
  • task/cancel

Primary event types:

  • task.started
  • task.progress
  • task.question
  • task.output
  • task.completed
  • task.failed
  • task.cancelled

Agent Behavior

  1. Run fix-officecli-env.sh before starting any bridge session so the skill bundle is refreshed and any missing officecli setup is repaired.
  2. Run check-officecli-env.sh after the repair step.
  3. Ensure officecli is installed, configured, and reachable.
  4. Ensure officecli agent-bridge can be started locally.
  5. Read initialize or capabilities/get before invoking generation, and cache document_generation.pptx.image_support and top-level image_generation.
  6. Also cache update; if available=true, use update_command or your own repair flow instead of parsing human CLI update prompts.
  7. Convert the user's natural-language request into:
    • document_type
    • topic
    • prompt
    • optional mode
    • optional lang
    • optional style
    • optional audience
    • optional ratio for img
  8. If the user explicitly wants no images for pptx, set enable_images=false; otherwise follow the bridge capability default instead of hard-coding a client default.
  9. For standalone img, call office.generate with document_type=img; do not call office.render, do not set mode=best, and do not use local image provider config.
  10. Use interactive=true by default so the chat can handle follow-up questions.
  11. Use mode=fast by default unless the user explicitly asks for a higher-quality, more iterative workflow and the document type is not img.
  12. On task.question, present the question naturally in the channel and forward the answer via task/respond.
  13. On task.output, read result.file_path and send the file as an attachment in the current channel.
  14. On task.failed, convert the error into a user-friendly message.
  15. On user cancel, send task/cancel.

Runtime Mode Rules

  • use officecli config runtime to inspect the local default runtime mode when the host asks how OfficeCLI is configured
  • use officecli config set-runtime hosted when the host explicitly wants platform-managed hosted generation by default
  • use officecli config set-runtime external when the host wants local/external generation by default
  • hosted mode requires a platform OfficeCLI API key with hosted credits; do not ask users for aigateway keys because those are created and stored by the platform
  • OfficeCLI should be installed through one channel at a time; if the host already has a Homebrew-installed officecli, do not suggest or run npm install -g officecli on top of it

PPT Image Rules

For all OpenClaw agents using this skill:

  • inspect document_generation.pptx.image_support.default_enabled during capability discovery
  • inspect update.available during capability discovery
  • use document_generation.pptx.image_support.disable_flag when explaining how to produce a text-only deck
  • if update.available=true, prefer a structured repair/refresh path and show update_command when the host asks how to update
  • use document_generation.pptx.image_support.config_command and config_fields when the user reports missing images
  • if task.output, task.completed, or task/status includes result_meta.image_support.attention_required=true, surface that immediately in the chat
  • if result_meta.image_support.reason=image_generation_degraded, tell the user the deck was downgraded to a no-image version and they should check image_base_url, image_api_key, and image_model
  • do not rely only on free-form warning strings for client decisions; prefer result_meta
  • do not parse human update prompts from officecli stdout; use bridge capability fields

Standalone Image Rules

For standalone img requests:

  • inspect top-level image_generation during capability discovery
  • use office.generate with document_type=img
  • pass ratio=square|landscape|portrait when the user asks for a shape; default to square
  • pass one reference_image local path or http/https URL when the user provides a reference image
  • require platform/license config and let the OfficeCLI server control the image provider
  • treat quota by runtime: in external mode one successful standalone image consumes one generation count, while in hosted mode it consumes hosted credits; free image usage has a separate 3-per-day bucket from free document generation
  • keep preview publishing enabled by default when publishing is configured; pass publish=false only for local-only output
  • do not use office.render, local config set-generation image settings, mode=best, source files, or local preview
  • include returned quota or credit balance metadata in the chat message when present

Environment Repair Rules

  • refresh the OpenClaw skill bundle and repair any missing officecli setup on every task by running fix-officecli-env.sh
  • do not refresh an already installed officecli binary unless the host explicitly opts in, for example with OFFICECLI_REFRESH_BINARY=1
  • when the user explicitly asks to uninstall officecli, run uninstall-officecli.sh
  • if the host previously installed with Homebrew and wants to switch to npm, tell them to run brew uninstall officecli/homebrew-officecli/officecli first; if their formula uses the short name, use brew uninstall officecli, then run npm install -g officecli
  • use check-officecli-env.sh as the single readiness probe for binary, config, and bridge
  • use fix-officecli-env.sh as the single repair entrypoint
  • when config is missing, ask only for the missing generation/license values and let the fix script write local config
  • online preview config is required by default so generated files can return publish URLs
  • if the current request is intentionally local-only, set OFFICECLI_SKIP_PUBLISH_SETUP=1 before running the fix script
  • do not try to start agent-bridge until the check script returns ready
  • if refresh or check fails, stop and report the officecli environment error; when repair fails, use the fix script's final structured status as the source of truth, and do not fall back to any other PPT/DOC/XLS generation tool without explicit user approval

Attachment Delivery

When generation succeeds:

  • read task.output.payload.result.file_path
  • upload that file to the current channel
  • include a short note with:
    • document type
    • document name
    • any warnings returned by bridge
    • when present, result_meta.image_support.message

Do not only send a local file path unless attachment upload is impossible on the current channel.

Conversation Policy

  • If document type is missing, ask which file type the user wants.
  • If topic or goal is missing, ask a concise clarifying question.
  • If the bridge emits task.question, relay it instead of inventing your own replacement question.
  • Keep progress updates short and stage-based.
  • do not trigger office.review / office.score automatically after generation unless the user explicitly asks for scoring, review, validation, or quality checking

Local Requirements

Expected local setup:

  • officecli available in PATH, or repairable by fix-officecli-env.sh
  • generation and license config already completed, or repairable by the fix script
  • standalone img requires license config; document/PPT image assets also require generation config
  • OpenClaw agent has permission to:
    • spawn local commands
    • read generated files
    • upload attachments to the active channel

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Openclaw Officecli AI skill do?

Use when an OpenClaw user clearly wants a local Office or image artifact such as a PPTX, DOCX, XLSX, Report, or IMG, and route the request through officecli agent-bridge instead of parsing human CLI output.

Why use Openclaw Officecli on TypingMind?

Because you install it once and use it with any model. Openclaw Officecli is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Openclaw Officecli in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/officecli/officecli/tree/main/plugins/openclaw-officecli/skills/openclaw-officecli. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Openclaw Officecli?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Openclaw Officecli?

As many as you like. As long as a model supports skills, you can use Openclaw Officecli with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Openclaw Officecli AI skill free?

Yes. It is published on GitHub by officecli under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇