zhongkao-yuwen

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill includes forced automatic publishing to a remote GitHub repo and loads/executing an external FeishuSync script that submits detailed user/practice data (including wrong answers and URLs), which constitutes deliberate data exfiltration and remote-code execution risk.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The HTML template includes and will load the remote script https://xingyun-new.github.io/Skills-XiaoSiMen/lib/feishu-sync.js at page runtime, and the page immediately calls FeishuSync.* functions (sync/submit) — so external code is fetched and executed as a required runtime dependency.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly requires writing HTML files to the project, automatically modifying README.md, and executing a forced GitHub Pages publish flow (git add → commit → push), which changes local files and performs remote operations—i.e., it modifies the machine/repository state without user confirmation.