Install
npx skills add WecomTeam/wecom-unified -y -gWho it is for
If you already handle enterprise office work inside an agent and need natural language to reach WeCom messages, mail, documents, todos, schedules, meetings, WeDrive, or the contact directory, this skill fits. Officially it is a WeCom CLI suite for raising office efficiency, not a second standalone console.
It names support for WorkBuddy, CodeBuddy, MiniMax Code, Kimi Work, Codex, Cursor and uses one command to auto-detect platforms already installed on this machine, then finish install. That only means those environments can be installed. It does not mean it can only work under those names, and it does not say other environments are excluded.
You must meet all of these: this machine is macOS（x64/arm64）, Linux（x64/arm64）, or Windows（x64）; Node.js is 18 or higher; and you already have a WeCom account. Without an account you cannot finish first-time credential setup. If the Node version is too low, even the install prerequisite fails.
A better fit can already send tasks in one of the platforms above and wants a read-only action first to confirm WeCom can return data, instead of blasting messages or changing a schedule immediately. Official notes are
https://github.com/wecomteam/wecom-unified
First try in 3 minutes
Use only this official install path. Do not try other installers in parallel:
npx skills add WecomTeam/wecom-unified -y -g
First confirm in a terminal that Node.js is 18 or higher, and that this machine already has a usable WeCom account. Then run the command above in an environment where a supported platform is already installed. Official notes say: one command auto-detects installed platforms and finishes install.
The install location is written by that command for the detected platform. The page does not give a manual copy directory. After install, the agent automatically gains WeCom operation ability. On first use the skill auto-checks and installs the needed CLI tools and walks you through credential setup. Before send, create, or upload, finish that credential walkthrough and confirm setup succeeded.
Pick a low-risk, checkable read for the first task: in a supported agent, ask in natural language for contact-directory member info, or to read an existing document, or to search documents. Do not start with bot-initiated notices, sending mail, creating todos, or uploading WeDrive files.
There are two observable success layers. First: the skill no longer stops on a missing CLI or unfinished credentials. Second: the agent returns contact-directory member info, or existing document content or search results, instead of reporting not installed, not configured, or no access.
Usage examples
Example 1: Read documents first, then decide whether to create.
Official text supports creating and reading documents, plus document search. Keep the first pass on read or search: have the agent open a known document, or search documents by keyword, and check whether the returned title and body snippet look like what you see in WeCom. Only after this step is stable, consider creating a document. The page gives no document id, space name, or CLI flags. What you can confirm stops at: these document actions can be triggered in natural language.
Example 2: Fetch meeting info first, then consider booking.
Official text supports creating and managing schedules, booking, and fetching meeting info. The first pass only fetches meeting info or views an existing schedule. After you confirm the returned time and topic, then talk about creating a schedule or booking a meeting. Likewise, the page writes no meeting-field template or booking arguments. The evidence only covers that these schedule and meeting actions exist, not a concrete form.
Both examples sit on the same install path and the first credential walkthrough. Sending mail, following up todos, and uploading WeDrive files are also in the capability inventory, but they change enterprise-side data and do not fit as a first try.
Limits
The runtime surface only writes macOS（x64/arm64）, Linux（x64/arm64）, and Windows（x64）, requires Node.js 18 or higher, and requires a WeCom account. Missing any one of these can stall install or first-time credentials.
Capabilities cover messages, mail, documents, todos, schedules, meetings, WeDrive, and the contact directory, and they name bot-initiated notices, upload and fetch of WeDrive files, send and fetch mail, and fetch of contact-directory member info. Write-class actions (send, create, book, upload, follow-up) change WeCom data. Confirm permission with read-class actions first.
The page does not write call quotas, visibility scope, or multi-enterprise switching, and it does not write a fallback path when you have no WeCom account. Officially it only guarantees that after install and credentials, the agent gains WeCom operation ability. Failure reasons beyond that need you to check local logs and account permissions.