Skill Security Scanner

自动检测并评分 ClawHub、GitHub 和本地 Skill 中 JS、TS、Python、Shell 文件的安全风险，识别数据泄露、注入、混淆和木马等威胁。

What to consider before installing or running this skill: - Purpose vs implementation: The skill's description and SKILL.md claim multi-language scanning and remote-URL support, but the Node scanner only walks .js files; confirm which script is actually used in your environment (the shell scripts appear to implement more checks). If you rely on .py or .sh scanning, verify the scanner actually inspects those files. - Remote targets: SKILL.md shows commands that accept GitHub/ClawHub URLs. Before passing a URL, inspect scan.sh to confirm it clones/downloads repositories in a safe, read-only way (and does not run any install scripts from the target). If unsure, clone the target manually into a sandbox and run the scanner against the local copy. - Execution risk: Running index.js will invoke the included shell scanner via execSync; running the bundled scripts will execute grep/sed/other command-line utilities and read many files. That is expected, but avoid running the scanner as root and run it in an isolated/sandboxed environment when scanning untrusted repos. - False positives and coverage: The scanner is rule-based with many aggressive regexes; expect false positives. The SKILL.md enforces a two-step flow (static report then optional LLM analysis) which helps avoid blind automated decisions — keep that workflow. - Verify truncations: Parts of scripts were truncated in the provided content. Review the full scripts/scan.sh and scan-all.sh to ensure there are no hidden fetch-or-execute steps (e.g., automatic 'curl | bash', running postinstall hooks, or executing target code). If you cannot inspect the full code, treat the skill as untrusted and run only in a sandbox. If you want, I can: (1) point out exact lines to inspect in scripts/scan.sh for remote-fetch behavior, (2) summarize where the node scanner and shell scanner differ, or (3) recommend a safe command sequence to run this scanner against a cloned repository in a disposable environment.

Type: OpenClaw Skill Name: skill-sec-scan Version: 4.4.1 The bundle is a security scanner tool designed to audit other OpenClaw skills for malicious patterns. It uses a combination of static analysis (via shell scripts and Node.js) and guided LLM semantic analysis to identify risks such as data exfiltration, command injection, and backdoors. The core logic in `scripts/scan.sh` and `node/scanner.js` uses regular expressions to flag dangerous code patterns (e.g., private keys, `eval`, `rm -rf`) and generates a detailed security report. While the tool performs high-risk operations like downloading remote content and executing shell commands, these actions are strictly aligned with its stated purpose as a security utility and are transparently documented in `SKILL.md` and `SKILL_EN.md`.