AI Agent Skill Scanner

对指定 Skill 目录做静态文本规则扫描，检出 eval/exec、管道到 shell、破坏性命令、硬编码密钥、凭证文件访问、base64 混淆等模式，输出控制台报告或 JSON。属规则匹配，无法识别混淆代码。

certainlogicai

热度 14.83

最近更新 2026-07-12 22:01:52

分类：安全依赖：代码执行 / 读写文件应用场景：代码审查 / 自动化脚本 / 工作流

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add ai-agent-skill-scanner

Metadata

name: AI Agent Skill Scanner
version: 2.1.0
description: Security scanner for AI agent skills. Detects hardcoded secrets, unsafe code execution, prompt injection, and malware patterns in under 50ms. Scan before you install.
metadata: {"openclaw":{"emoji":"🔒","tags":["security","scanner","malware-detection","prompt-injection","secrets-detection","code-scanning","skill-safety"]}}

AI Agent Skill Scanner 🔒

Scan any skill for security issues before you install it. Detects hardcoded secrets, unsafe code execution, prompt injection, and malware patterns in under 50ms.

How to Use

# Scan any installed skill folder
python3 scripts/vetter.py /path/to/skill

# JSON output for scripting
python3 scripts/vetter.py /path/to/skill --json

What It Detects

Issue	Severity	Example
`eval()` / `exec()`	Critical	Runs arbitrary code
Pipe to shell	Critical	`curl ...
Destructive commands	Critical	`rm -rf /`, `mkfs.`
`os.system()`	Critical	Shell commands
Hardcoded API keys	High	`api_key`, `apikey`
`shell=True`	High	Shell injection risk
Hardcoded secrets	High	`auth_token`, `secret`
Credential file access	High	`.ssh/`, `.aws/`
Sudo elevation	High	`sudo` commands
Base64 obfuscation	High	Hidden code execution
Git config access	High	`.git/config` reads
Password hardcoding	High	`password` literal
Permission overrides	Medium	`chmod 777`
Network calls	Medium	`requests.post()`

New in v2.1

10 new detection signatures (pipe-to-shell, sudo escalation, destructive commands, credential file access, base64 obfuscation, git config, permission overrides)
Skips node_modules/, __pycache__/, .git/ (reduced false positives)
Smarter pipe-to-shell detection (only flags actual pipe patterns)
Self-scan exclusions (scanner ignores its own signatures file)
--verbose flag for skipped file counts

Example Result

Scanned 430 files in 368ms (skipped 6)
Found 1 issue(s):
  [CRITICAL] pipe-to-shell at install.sh:26
    → Remote code execution risk: downloading and piping to shell

Install

clawhub install ai-agent-skill-scanner

Notes

Text matching — won't catch obfuscated code
Always scan skills from unknown publishers before installing
Quick scan prevents accidents (50ms per skill)

Feedback

clawhub star ai-agent-skill-scanner — clawhub sync