ruanzhu

[Skill Scanner] Installation of third-party script detected The skill's stated purpose (generate a source-code DOCX) is plausible and the provided scanning/rendering rules are coherent. However, the skill mandates copying and executing an opaque Python script from ~/.claude/templates/ruanzhu/generate_docx.py and deleting it afterward, without any integrity checks or included source in the skill. This is a significant supply-chain and execution-risk pattern: the script can read all project files, environment variables, install packages, and perform network I/O (exfiltration) without visibility. There are internal inconsistencies (forbidding creating .py files or installing deps vs. error handling stating pip install will occur). Verdict: SUSPICIOUS. The package/text does not itself contain visible malware, but executing the opaque script carries high risk. Recommend not running the enforced command unless the template script has been inspected and its provenance verified. LLM verification: This skill's stated goal (generate a source-code DOCX) is plausible, but the execution model is high-risk and opaque. Forcing execution of a Python script copied from a writable local template directory (~/.claude/templates/ruanzhu/generate_docx.py), combined with runtime pip installs and prohibitions on inspection, creates a credible possibility for misuse (arbitrary code execution, data exfiltration, dependency-based supply-chain attacks). I classify this as suspicious: use only after verifyin