Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
bozo-wechat-publisher
v1.1.0一键发布 Markdown 文章到微信公众号草稿箱。当用户提到发布到微信、公众号、推文、草稿箱、上传文章时触发。支持 wenyan-cli 完整排版和 curl 备用方案,兼容所有 Node.js 版本。
⭐ 0· 24·0 current·0 all-time
by@bozoyan
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Functionality matches the description: scripts render Markdown (wenyan-cli) and call WeChat APIs to upload images and create drafts. However the package metadata declares no required env vars or config paths, while the runtime instructions and scripts clearly require WECHAT_APP_ID and WECHAT_APP_SECRET (and suggest persisting them to shell rc files). This mismatch (declared requirements: none vs actual: WeChat creds required) is an incoherence.
Instruction Scope
SKILL.md and bundled scripts instruct the agent/user to read credentials from ~/.openclaw/workspace/TOOLS.md (scripts/setup.sh, publish*.sh) and to run wrapper/install commands that modify user PATH. Reading $HOME/.openclaw/workspace/TOOLS.md is scope creep relative to a simple 'publish' skill because it accesses an arbitrary, user-local file that may contain other secrets; while the scripts parse only WECHAT_APP_ID/WECHAT_APP_SECRET, the act of reading that workspace file is not declared in metadata and could expose unexpected data if misused. The scripts also run network calls to api.weixin.qq.com (expected for publishing) and curl ifconfig.me (to get public IP for whitelist) — those network calls align with purpose but the TOOLS.md dependency is unexpected.
Install Mechanism
There is no install spec (instruction-only), and all code is local in the skill bundle. No remote downloads or extracted archives are declared in an install step. The README instructs installing wenyan-cli from npm (a normal dependency) and creating a local wrapper script — this is a standard workaround but it modifies user's PATH and writes to ~/.local/bin if followed.
Credentials
The skill requires WECHAT_APP_ID and WECHAT_APP_SECRET at runtime but the registry metadata listed no required env vars or primary credential. The scripts also implicitly require access to a user-local TOOLS.md at $HOME/.openclaw/workspace/TOOLS.md and will grep it for credential exports. Requesting/reading those files and env vars is reasonable for WeChat publishing, but the omission in declared requirements and the use of an undeclared workspace file are disproportionate and should be clarified before install.
Persistence & Privilege
The skill is not marked always:true and does not request elevated platform privileges. It suggests adding environment variables to shell rc files and creating a local wrapper script (user-performed actions), which are normal for CLI tooling but are not automatic in the bundle. The skill does not modify other skills or system-wide agent settings in the code provided.
What to consider before installing
This skill appears to implement the promised WeChat publishing flow, but there are a few mismatches you should consider before installing:
- The registry metadata lists no required env vars, yet SKILL.md and the scripts require WECHAT_APP_ID and WECHAT_APP_SECRET. Decide whether you are comfortable providing those credentials to scripts you run locally.
- The scripts try to load credentials from ~/.openclaw/workspace/TOOLS.md. Inspect that file yourself first — if it contains other secrets, avoid letting scripts read it. Prefer explicitly exporting only the two WECHAT_* env vars in a dedicated file you control, or set them in your shell session before running.
- The install instructions advise creating a wrapper in ~/.local/bin and appending to your shell rc. Those are user-level changes; review and run them manually rather than blindly executing the provided snippets.
- The scripts upload images and send HTML content to api.weixin.qq.com (expected). If you keep sensitive content in your Markdown or assets, be aware they will be transmitted to WeChat.
- If you want to reduce risk: run the publish scripts in an isolated environment (container or dedicated account), or audit/modify the scripts to remove the TOOLS.md lookup and to print the API payload for review before sending.
If you want, I can: (a) point out the exact lines in scripts that read TOOLS.md and network endpoints, (b) produce a minimal patched version that avoids reading TOOLS.md, or (c) generate a checklist to run this safely in a container.Like a lobster shell, security has layers — review code before you run it.
latestvk973revfynfefrjtxezydjdspn8477pn
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
📱 Clawdis
