Install
npx skills add WeComTeam/wecom-cli -y -gWho it's for
Use this skill when you want to operate WeCom from the terminal — messages, mail, documents, todos, calendar, meetings, WeDrive, or contacts — and you want humans and an AI Agent to share the same command line. It is not another web admin; it brings open-platform capability to the terminal to raise office efficiency. Official repository:
https://github.com/wecomteam/wecom-cli
Preconditions are explicit: this machine needs Node.js >= 18 and a WeCom account; platforms are limited to macOS (x64/arm64), Linux (x64/arm64), and Windows (x64). Without an account, if the version is too low, or if you are off those platforms, do not install yet. People who only want to click WeCom in a browser and do not want a global command on this machine should not start here.
The source splits CLI install and CLI Skill install into two steps and writes that the Skill is required. Have the command line first, then let the Agent learn to call it. Do not invent an installer that never appeared, and do not add an unnamed skill directory. The source does not write “humans and AI Agent” as exclusive to a few products; do not narrow the named clients yourself.
First try in 3 minutes
First confirm three things: Node.js is 18 or higher; you have a WeCom account; the current system is among the platforms above. Open the official page and check:
https://github.com/wecomteam/wecom-cli
The main path is install the CLI first, then the required CLI Skill. Installing the package changes this machine's global commands; before you run it, confirm you want @wecom/cli left on this machine:
npm install -g @wecom/cli
After that, run the Skill install the source gives (marked required):
npx skills add WeComTeam/wecom-cli -y -g
Do not split these two steps into two parallel main paths, and do not change the parameters. The source does not write the skill landing directory; do not invent a path.
The first time in an Agent, only check; do not create documents, send mail, or change calendar. If credentials are not configured yet, configuration is a side effect: QR scan by default, or --manual to type. Before you run it, confirm this is your own WeCom account and you are ready for this interaction:
wecom-cli auth init
After configuration, inspect authorization with a read-only command:
wecom-cli auth show
Success signal: the command ends normally and you can see the current authorization state. The source gives no field names for show; do not invent an output format. The first time, do not do bot notifications, document writes, meeting booking, or WeDrive uploads. If the Agent can only recite “init first, then show” and has not actually run show, do not treat authorization as already successful.
Use cases
Case 1: Only check authorization. After CLI and the required Skill are installed, run wecom-cli auth show in the terminal or in an Agent that can call the same commands. This inspects authorization state and is for confirming credentials are still there. Before you run it, confirm you will not change configuration along the way; the success signal is seeing the authorization state. The source does not write an exit code for unauthorized; on failure, stop, do not invent a repair command. Do not expand this check into sending mail or reading contacts; those actions have no corresponding command wording.
Case 2: Configure credentials the first time. The source writes that configuring credentials is interactive and needed only once, QR scan by default, --manual to type. The command is wecom-cli auth init. Before you run it, confirm you want WeCom credentials left on this machine and are ready to scan or type. After configuration, run wecom-cli auth show to check. The source writes no revoke or sign-out command; do not invent a credential-uninstall step. Business capabilities — messages, documents, todos, calendar, meetings, WeDrive, mail, contacts — are written only as capability names, for example create and read documents, document search, create and manage calendar, book and fetch meeting info, create and follow todos, upload and fetch WeDrive files, send and fetch mail, fetch contact member info, and bot-initiated notifications; when you need those operations, go back to the official repository first, do not invent commands.
Boundaries
Platform, runtime, and account requirements follow the source: macOS (x64/arm64), Linux (x64/arm64), and Windows (x64), Node.js >= 18, and a WeCom account. Humans and an AI Agent can both operate in the terminal, but the source names no specific Agent product; do not write a few clients as exclusive.
Install order is npm install -g @wecom/cli, then the required npx skills add WeComTeam/wecom-cli -y -g. Do not invent a second marketplace or an unwritten skills add variant. Do not read “required Skill” as permission to skip the CLI; both steps must finish in the source order.
Credential configuration writes to this machine, QR scan by default, --manual available. The source does not write the credential file path, expiry policy, or how to revoke. Business coverage includes messages, mail, documents, todos, calendar, meetings, WeDrive, contacts, including bot-initiated notifications, document search, booking meetings, WeDrive upload, but evidence stops at capability description and gives no wording for those subcommands; do not fill in unwritten parameters, permission scopes, or billing. The first time, only verify authorization; do not treat office actions without commands as already copyable steps. Do not invent uninstall or update paths the evidence never wrote.