Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed WeChat-to-Obsidian clipping workflow, with limited local file writes that match its stated purpose, though the package appears incomplete.
Before installing, be aware that this skill is intended to fetch WeChat article content and write Markdown/images into a local Obsidian vault or chosen directory. Confirm the output path before use, use --no-img if you do not want images downloaded, and verify the missing script/config/requirements files are included in any actual install package before expecting it to run.
The skill description includes broad natural-language trigger phrases like '存一下 / 保存到笔记 / 转成 Markdown / 剪藏到 Obsidian / 保存到知识库', which can match ordinary conversation too loosely when a WeChat URL is present. This increases the chance of unintended invocation, causing the agent to fetch remote content and write files into a local Obsidian vault without sufficiently explicit user intent.
The trigger condition says the skill should activate when the user provides a WeChat article link and asks to save/clip/convert it, but it does not define clear non-trigger cases or disambiguation rules. In an agentic environment, ambiguous activation can lead to unnecessary network requests and local filesystem writes, which are meaningful side effects even if the content source is limited to mp.weixin.qq.com.
No suspicious patterns detected.
