Weixin Send Media

微信发图片/文件技能 - 解决 contextToken 持久化问题

This skill does what it claims — enabling CLI/scripted Weixin media sends by persisting context tokens — but it patches OpenClaw's weixin gateway code and writes sensitive tokens to disk in plain JSON. Before installing: (1) inspect the actual patch file (patches/inbound.ts.patch) line-by-line to ensure no unexpected behavior, (2) back up ~/.openclaw/extensions/openclaw-weixin/src/messaging/inbound.ts (the installer does this, but verify the backup), (3) consider running the install in a disposable environment (VM/container) first, (4) if you accept installation, restrict token files (chmod 600) and consider adding encryption-at-rest or a more secure storage mechanism, (5) review export-context-token.js to avoid accidentally printing full tokens to logs or stdout, and (6) only install if you trust the author/repository — otherwise prefer an alternative solution that doesn't require patching core files. If you want higher confidence, provide the exact inbound.ts.patch content for a focused audit.

Type: OpenClaw Skill Name: weixin-send-media Version: 1.1.0 The skill modifies OpenClaw core extension files (~/.openclaw/extensions/openclaw-weixin/src/messaging/inbound.ts) via a patch to persist WeChat contextTokens to the local filesystem (~/.openclaw/openclaw-weixin/context-tokens/). While the documentation in SKILL.md and the logic in install.sh are transparent and aligned with the stated purpose of enabling CLI-based media sending, the practice of patching core system files and storing session tokens on disk represents a high-risk capability that could be leveraged for unauthorized access or system instability.