Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed builder for creating an Obsidian article-organizer skill, with local file movement risks that are described and bounded rather than hidden.
Before installing, be aware that using this builder will create a persistent local skill and that the generated organizer may move processed source articles and attachment folders into an archive or recycle location. Review the generated paths and cleanup behavior before running the generated skill on important notes.
Referenced artifact was not completely inspected
3. **生成目标 skill**：在 `~/.workbuddy/skills/<用户主题相关的名字>/` 下按规范 8 步生成完整目录：`SKILL.md` + `scripts/scan.py` + `scripts/cleanup.py` + `references/classification.md`。规范中
The description advertises 'cleanup source files' behavior but does not clearly warn about the data-impact risks or require explicit approval for destructive or quasi-destructive actions. Even though the note mentions 'archive not delete' for generated scripts, the skill's own high-level description still normalizes source-file cleanup without defining safeguards, which can lead to accidental data loss or unexpected file state changes.
The skill explicitly states it will generate and write a new skill under ~/.workbuddy/skills/, but it does not require explicit user confirmation immediately before performing that filesystem modification. Because this is a meta-skill that creates executable automation artifacts, silent or implicit writes could surprise the user, create persistence, or install unsafe logic into the user's environment.
The skill goal states that after archiving, the source files will be cleaned up automatically, which affects user data and file location. Although later sections describe safety constraints like archiving or recycle-bin use, the description does not clearly warn the user up front that successful execution will move their original files and attachments out of the source directory.
The instruction "用中文写一份流程说明" forces a specific language for generated skill documentation. There is no indication that the user may opt into another language or that the skill is region-specific enough to justify a mandatory Chinese-only policy.
No suspicious patterns detected.