← 返回 Skills 市场
mmxagent-skill-wechat
作者
OreoAndYuumi
· GitHub ↗
· v1.0.3
· MIT-0
120
总下载
0
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install mmxagent-skill-wechat
功能描述
连接个人微信（不是企业微信）。用户说"连接个人微信"、"接入个人微信"、"绑定个人微信"、"个人微信扫码"时使用本 skill。注意：如果用户说的是"企业微信"或"企微"，本 skill 不适用，请使用 wecom-connect skill。一旦匹配本 skill，必须严格按流程执行到底，不得跳步或自由发挥。
安全使用建议
This skill appears to do what it says: install/upgrade OpenClaw and a Tencent WeChat plugin, fetch a one-time login QR, wait for you to scan/confirm, then save the returned bot token into ~/.openclaw/openclaw-weixin/accounts. Before running it: (1) confirm you trust the npm packages and the registry mirror (registry.npmmirror.com and @tencent-weixin/openclaw-weixin-cli), (2) be aware the flow will run npm install -g openclaw which modifies your environment, (3) do not use this in untrusted/shared or ephemeral remote shells unless you accept the token being written to your HOME, and (4) avoid using the optional CDN upload unless you understand it will expose the one-time QR to an external host. Note that the provided SKILL.md appears truncated at the end — verify the full skill text (especially final write/restart commands) before running commands it suggests.
能力评估
Purpose & Capability
The skill's declared purpose (connect personal WeChat) matches the actions it instructs: installing an OpenClaw WeChat plugin, requesting a WeChat QR via ilinkai.weixin.qq.com, generating a QR image, polling for scan/confirmation, and writing the returned bot token/account file into ~/.openclaw/openclaw-weixin/accounts. No unrelated services or credentials are requested.
Instruction Scope
Instructions are narrowly focused on the connection flow and explicitly forbid reading other docs or auto-polling. They do require running system commands (openclaw plugin list/uninstall, npm/npx installs, curl) and writing credential files to the user's home. The skill requires user confirmation before writing sensitive tokens, which is appropriate. Note: the SKILL.md provided here appears truncated at the end (the write script is cut off), so final cleanup/restart steps are not visible in the supplied content.
Install Mechanism
This is an instruction-only skill (no install spec), but it instructs runtime installation of npm packages: global openclaw (npm install -g from registry.npmmirror.com), the Tencent plugin via npx, and qrcode in /tmp. Using a registry mirror and npx is common, but runtime installs execute third-party code on the host and are a moderate-risk action; users should confirm trust in these packages and the registry.
Credentials
The skill requests no environment variables or unrelated credentials. It does write a bot token and metadata into ~/.openclaw/openclaw-weixin/accounts/<accountId>.json — which is necessary for the stated purpose. The skill mandates showing a masked token to the user and requiring explicit confirmation before writing, which is proportionate.
Persistence & Privilege
always:false and model invocation are normal. The skill persists credentials to ~/.openclaw (its own config area), which is expected. It does not request system-wide privileges or modify other skills' configs. Users should be aware that installing/upgrading openclaw globally will alter their environment.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install mmxagent-skill-wechat - 安装完成后，直接呼叫该 Skill 的名称或使用
/mmxagent-skill-wechat触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.3
mmxagent-skill-wechat v1.0.3
- 二维码相关步骤（获取+生成）明确要求用命令合并执行，以提高速度，降低过期风险。其它步骤允许分步操作便于排错。
- 二维码图片展现方式调整为优先存储本地 workspace 文件，CDN 上传仅作为辅助，需用户主动请求时才启用。
- 新增凭证写入前，先展示账号ID/token等敏感信息，获得用户确认后再落盘，增强安全与可控性。
- 其他部分进一步细化执行流程，明示各步骤目的与合规要求，加强对包来源、流程禁止项的说明。
- 绝对禁止事项补充完善，CDN 上传、凭证写入等关键环节限制更严格。
v1.0.2
mmxagent-skill-wechat v1.0.2
- 增加了对旧版 openclaw(<2026.3.13) 环境需先删除旧微信插件再安装新插件的流程，避免版本冲突。
- 路径 A 现在明确区分：检测是否有旧版插件，若存在先卸载，再进行新插件安装和 openclaw 升级。
- 新增卸载旧插件的检测与操作说明，保证插件安装的兼容性和稳定性。
- 其他执行流程、二维码展示、凭证写入等步骤保持不变。
v1.0.1
**重要更新：新增 openclaw 版本检测步骤，流程更适配不同 openclaw 版本。**
- 新增“第 0 步”：上线前检测 openclaw 版本，分支处理插件与 openclaw 升级序列。
- 安装流程分为旧版本（≤2026.3.13）和新版本（>2026.3.13）两条分支，分别处理插件和 openclaw 升级顺序。
- 新增对未安装 openclaw 情况的提示与处理。
- 强化绝对禁止事项，对跳过“第 0 步”、终端字符渲染二维码等方式做进一步限制说明。
- 其余执行流程保持不变，保证二维码处理与用户引导一致。
v1.0.0
- Initial release of mmxagent-skill-wechat for connecting personal WeChat (not WeCom).
- Strict matching: only triggers for personal WeChat keywords; ignores WeCom/企业微信/企微 related requests.
- Follows a precise step-by-step workflow: install legacy plugin, fetch QR code, generate/upload QR image, guide user, poll for status, write credentials, and restart gateway.
- Includes strict rules on command merging, QR code generation (must use qrcode npm library), and file upload (only upload_to_cdn or workspace file allowed).
- Strong error handling for failed steps and QR code expiry, with clear guidance for users.
- Explicitly forbids alternative behaviors (e.g., using original QR code URL, auto polling, or unsupported upload/display methods).
元数据
常见问题
mmxagent-skill-wechat 是什么？
连接个人微信（不是企业微信）。用户说"连接个人微信"、"接入个人微信"、"绑定个人微信"、"个人微信扫码"时使用本 skill。注意：如果用户说的是"企业微信"或"企微"，本 skill 不适用，请使用 wecom-connect skill。一旦匹配本 skill，必须严格按流程执行到底，不得跳步或自由发挥。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 120 次。
如何安装 mmxagent-skill-wechat？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install mmxagent-skill-wechat」即可一键安装，无需额外配置。
mmxagent-skill-wechat 是免费的吗？
是的，mmxagent-skill-wechat 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
mmxagent-skill-wechat 支持哪些平台？
mmxagent-skill-wechat 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 mmxagent-skill-wechat？
由 OreoAndYuumi（@oreoandyuumi）开发并维护，当前版本 v1.0.3。
推荐 Skills