← 返回 Skills 市场
900
总下载
1
收藏
9
当前安装
2
版本数
在 OpenClaw 中安装
/install wecom-doc
功能描述
文档与智能表格操作。当用户提到企业微信文档、创建文档、编辑文档、新建文档、写文档、智能表格时激活。支持文档创建/写入和智能表格的创建及子表/字段/记录写入。注意：所有文档创建和编辑请求都应使用此 skill，不要尝试用其他方式处理文档操作。
安全使用建议
Before installing or enabling this skill, consider the following: (1) Confirm the origin and trustworthiness of the 'mcporter' npm package (review its publisher, repository, and code) because the skill installs it globally. (2) Inspect the file ~/.openclaw/wecomConfig/config.json yourself to see what data it contains—it may hold connection URLs or tokens the skill will read and use. (3) Ask the skill author why metadata omits the config path and why always:true is required; prefer a variant that requires explicit user enablement. (4) If you proceed, avoid granting broad, always-on privileges: disable always:true if possible, and run initial installs in a sandbox or controlled environment. (5) Require explicit user confirmation before the agent performs npm install -g and before it transmits any data to remote endpoints. If you can get the skill source (homepage/repo) or more info about mcporter and the MCP server, share those for a higher-confidence assessment.
功能分析
Type: OpenClaw Skill
Name: wecom-doc
Version: 1.0.1
The wecom-doc skill bundle provides comprehensive tools for managing WeCom (Enterprise WeChat) documents and smart sheets via the mcporter MCP bridge. It includes automated environment checks, dependency installation (npm install -g mcporter), and configuration management logic that reads from ~/.openclaw/wecomConfig/config.json. While it utilizes shell execution and local file access, these actions are strictly aligned with its stated purpose of integrating WeCom document services into the OpenClaw environment and use legitimate WeCom domains (work.weixin.qq.com) for authorization.
能力评估
Purpose & Capability
The skill's name and description (WeCom document / smart-sheet operations) align with using a local mcporter CLI to call an MCP server. However the runtime instructions require reading ~/.openclaw/wecomConfig/config.json to auto-configure mcporter (extracting mcpConfig.doc.type and url), but the skill metadata declared no required config paths or credentials. That mismatch (metadata says no config paths but SKILL.md reads a specific user file) is an incoherence.
Instruction Scope
The SKILL.md instructs the agent to run shell commands (which mcporter, mcporter list, npm install -g mcporter, cat ~/.openclaw/wecomConfig/config.json, mcporter config add, mcporter call ...) and to automatically continue after installs/configs. Reading the user's ~/.openclaw/wecomConfig/config.json is outside what the registry metadata declared and may expose connection details or secrets; the skill also mandates automatic, seamless continuation after installation/config steps, giving it broad runtime behavior beyond a single explicit user action.
Install Mechanism
Install is via npm package 'mcporter' (node). Using an npm package for a CLI is expected for this use case, but installing globally with 'npm install -g' writes binaries to the system and pulls code from the public registry—this is moderate-risk and should be vetted (confirm package provenance and contents). The install URL/host is a standard package registry (not an arbitrary URL), which is better than downloading an archive.
Credentials
The skill declares no required environment variables or credentials, yet its instructions read a local config file that likely contains MCP connection parameters (mcpConfig.doc.{type,url}) and then call mcporter to add that server. The skill therefore accesses configuration data that is not declared in metadata; requesting/using that file may expose sensitive endpoints or secrets without explicit declaration. The number of required secrets is not high, but the undeclared config access is disproportionate to the metadata.
Persistence & Privilege
always: true is set in the skill metadata. That forces the skill to be present/eligible in every agent run; combined with the ability to install a global CLI and read a user config file, this increases potential blast radius. The SKILL.md does not present a clear justification for always:true (it claims to activate when user mentions doc tasks), so this is unnecessary privilege.
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 wecom-doc - 安装完成后，直接呼叫该 Skill 的名称或使用
/wecom-doc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
v1.0.1
wecom-doc 1.0.1 更新日志
- 支持自动读取 wecom channel 长连接写入的本地配置文件（~/.openclaw/wecomConfig/config.json），在未配置 MCP Server 时自动完成 wecom-doc mcp server 配置，提高首次使用的便捷性
- 智能表格字段及数据写入要求调整：添加记录时 key 必须使用字段标题 (field_title)，且部分字段值（如日期）格式要求对齐企业微信侧新版要求
- fieldType、cellValue等映射、示例和易错点描述全面校正，确保与实际协议一致
- 优化和明确 MCP 错误响应处理逻辑，区分 help_message、错误码等不同场景的用户提示
- 文档细节调整：不再强调“仅处理企业微信文档”，扩展为所有文档创建/编辑统一交由本 skill，降低使用门槛
v1.0.0
wecom-doc 1.0.0
- 新增企业微信文档与智能表格一体化操作能力，支持文档和智能表格的全流程创建及写入。
- 默认用于所有未指定平台的“创建文档”“写文档”等相关请求，无需用户主动选择平台。
- 严格 docid 管理，仅允许操作由本 skill 创建的文档/表格，不支持通过 URL 编辑。
- 引入 mcporter 依赖及自动化安装/配置校验流程，并提供详细环境异常和权限引导提示。
- 内置表格字段类型与数据格式规则，确保数据写入合规。
- 业务错误支持 markdown 透传，提升问题排查和用户体验。
元数据
常见问题
wecom-doc 是什么？
文档与智能表格操作。当用户提到企业微信文档、创建文档、编辑文档、新建文档、写文档、智能表格时激活。支持文档创建/写入和智能表格的创建及子表/字段/记录写入。注意：所有文档创建和编辑请求都应使用此 skill，不要尝试用其他方式处理文档操作。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 900 次。
如何安装 wecom-doc？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install wecom-doc」即可一键安装，无需额外配置。
wecom-doc 是免费的吗？
是的，wecom-doc 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
wecom-doc 支持哪些平台？
wecom-doc 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 wecom-doc？
由 wecom（@garfileds）开发并维护，当前版本 v1.0.1。
推荐 Skills