Wecom Deep Op

基于微信官方插件 @wecom/wecom-openclaw-plugin v1.0.13+，封装的一站式企业微信自动化解决方案 - 你可以方便操作文档、日历、会议、待办、通讯录所有企业微信MCP能力，充分发挥OpenClaw与企业微信的协同能力。

This skill appears to implement a legitimate Enterprise WeChat (WeCom) MCP wrapper and needs the bot's uaKey (supplied by you) and local OpenClaw config to work — that is expected. However: - Metadata mismatch: the registry metadata says no env vars/config paths are required, but SKILL.md/README require WECOM_*_BASE_URL environment variables or an mcporter.json entry (uaKey). Confirm the published skill.yml actually declares the env vars before installing. - Audit before running anything that executes shell commands: the repo contains maintainer scripts (publish, build) that call execSync/grep/npm; do not run those unless you trust the source. They are not required for normal runtime but could be misused if executed. - Inspect runtime code (src/index.ts / dist/) for any network calls beyond the user-configured WeCom endpoints (qyapi.weixin.qq.com). Confirm logging does not print full URLs or uaKey. The docs claim uaKey is never stored, but verify in code. - Keep uaKey secret: treat it like a password. Configure it via env vars or mcporter.json with appropriate file permissions. Consider testing in a sandboxed environment or on a non-production account first. - If you need higher assurance: request the published package (skill.yml and dist) from the maintainer or from Clawhub, verify skill.yml lists the env vars and inspect the bundled dist for unexpected outbound endpoints before installing.

Type: OpenClaw Skill Name: wecom-deep-op Version: 2.0.2 The wecom-deep-op skill is a comprehensive and well-documented wrapper for Enterprise WeChat (WeCom) MCP capabilities, including document, schedule, meeting, todo, and contact management. The source code (src/index.ts) unifies access to these services by proxying requests to official WeCom endpoints (qyapi.weixin.qq.com) using user-provided environment variables or configuration files. The bundle includes robust input validation,指数退避 (exponential backoff) retry logic, and extensive security documentation (SECURITY_AUDIT.md and README.md) that explicitly addresses credential safety and data flow boundaries. No evidence of malicious intent, unauthorized data exfiltration, or prompt injection was found.