Token导航 LogoToken导航

Agent Skill · 研究检索

security-audit(安全审计)

用于辅助安全审计、权限检查、凭据风险、认证流程和常见漏洞排查。它适合让 Agent 梳理敏感配置、检查依赖风险、分析鉴权逻辑或生成安全复核清单。使用时不能把工具输出直接当最终结论,涉及密钥、令牌、用户数据或生产系统时,应先确认最小权限、脱敏方式和操作边界。

security-audit 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 64,698。

总安装 294周安装 12GitHub Stars 4下载 94核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:security-audit(安全审计)
来源仓库:https://github.com/hexters/laravel-security-audit
仓库路径:skills/security-audit
安装命令:
npx skills add https://github.com/hexters/laravel-security-audit --skill security-audit
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/hexters/laravel-security-audit --skill security-audit

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于辅助安全审计和漏洞排查。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。适合梳理敏感配置、检查依赖风险或分析鉴权逻辑。通过 npx skills add 命令从指定 GitHub 仓库安装使用。涉及密钥或用户数据时应先确认最小权限和操作边界。

security-audit 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/security-audit

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/hexters/laravel-security-audit --skill securit…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

security-audit 怎么安装?

在终端执行 npx skills add https://github.com/hexters/laravel-security-audit --skill security-audit。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

security-audit 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

security-audit 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

security-audit 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

security-audit 用的人多吗?

公开数据是累计安装 294 次,近一周安装 12 次,下载 94 次,GitHub 4 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

security-audit 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
security-audit

安全审计 · 本页

2944接触敏感数据未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Security Audit Skill for Laravel/PHP Projects

You are a security audit agent. When activated, perform a comprehensive security audit on the current project. Follow each phase in order and report findings clearly.

Phase 1: Vulnerable Package Detection

Scan composer.lock for known vulnerable packages:

# Check for intervention/image vulnerability (path traversal CVE)
python3 -c "
import json, sys
try:
    with open('composer.lock') as f:
        data = json.load(f)
    vulns = []
    checks = {
        'intervention/image': lambda v: int(v.lstrip('v').split('.')[0]) < 3,
        'symfony/http-kernel': lambda v: True,  # always check advisory
        'laravel/framework': lambda v: True,
    }
    for p in data.get('packages', []):
        if p['name'] in checks:
            ver = p['version'].lstrip('v')
            vulns.append(f\"{p['name']}: {ver}\")
    for v in vulns:
        print(v)
except FileNotFoundError:
    print('No composer.lock found')
"

Also run composer audit if available.

Action items:

  • Flag intervention/image < v3 as CRITICAL (path traversal, allows reading.env)
  • Check if vulnerable packages are used directly or as transitive dependency
  • Determine if upgrade is safe or needs local path repository approach

Phase 2: Malware & Webshell Scanner

Search for malicious files across these locations:

2a. PHP files in upload/storage directories (should NOT exist)

find storage/app/public public/uploads public/images public/storage -name "*.php" -type f 2>/dev/null

2b. Suspicious file patterns

Search for files with:

  • Random 5-character names: [a-z0-9]{5}.php
  • Known malware names: stmept, guard.sh, wp-admin.php, wp-cron.php, cache.php, test.php, unzip.php, bootstrap.cache.php, themes.php, menu.php, php8.php
  • Binary executables in web directories
  • Files with double extensions: .php.jpg, .jpg.php
  • Temporary PHP files: tmp_*.php

2c. Dangerous code patterns in PHP files

Search app/ and public/ directories for:

eval(
base64_decode(
exec(
system(
passthru(
shell_exec(
gzinflate(
str_rot13(
assert(
preg_replace with /e modifier
$_GET/$_POST/$_REQUEST used with eval/exec
create_function(

2d. Crypto miner detection

# Check for known miner processes
pgrep -f "xmrig\|stmept\|minerd\|cpuminer" 2>/dev/null

# Check for miner binaries
find . -name "stmept" -o -name "xmrig" -o -name "minerd" 2>/dev/null

# Check for mining pool connections in code
grep -r "supportxmr\|nanopool\|minergate\|coinhive" . --include="*.php" --include="*.sh" --include="*.conf" 2>/dev/null

2e. Injected.htaccess files

# Check all .htaccess files for injected content (nginx configs, shell commands, redirects)
find . -name ".htaccess" -exec grep -l "sudo\|nginx\|server_name\|proxy_pass\|RewriteRule.*http" {} \;

Phase 3: Configuration Security Check

3a. Storage directory protection

Check if storage/app/public/.htaccess exists with PHP execution blocking:

<FilesMatch "\.(?:php|phtml|phar|php[0-9]|inc)$">
    Order Deny,Allow
    Deny from all
</FilesMatch>

If missing, CREATE IT - this is the #1 defense against uploaded webshells.

3b. Public.htaccess hardening

Check public/.htaccess for:

  • Block PHP execution in storage/uploads dirs: RewriteRule ^storage/.*\.php$ - [F,L]
  • Security headers: X-Content-Type-Options, X-Frame-Options, X-XSS-Protection
  • Block sensitive file access: .env, .log, .sql, .bak

3c. File upload security

Search for these risky patterns in Filament/Controller code:

  • editableSvgs(true) - SVG can contain JavaScript (XSS) and XXE attacks
  • preserveFilenames() - Original filenames can be manipulated for path traversal
  • Missing mime type validation on upload fields
  • Missing file extension validation

3d. Credential exposure

Check .env file types present (DO NOT output actual values):

grep -E "^[A-Z_]+=." .env | sed 's/=.*/=***/' | grep -iE "key|secret|password|token|mail|sendgrid|aws|stripe|paypal"

Phase 4: Server-Level Checks (if accessible)

4a. Web server user

ps aux | grep -E "nginx|php-fpm|apache" | head -5

CRITICAL if running as root - web server must run as unprivileged user (www-data/nginx).

4b. Sudoers audit

grep -i "NOPASSWD" /etc/sudoers /etc/sudoers.d/* 2>/dev/null

Flag any NOPASSWD: ALL entries - especially for www-data, nginx, or CI users like gitlab-runner.

4c. Crontab persistence check

crontab -l 2>/dev/null
crontab -u www-data -l 2>/dev/null
ls -la /etc/cron.d/ /etc/cron.daily/ 2>/dev/null

4d. Active miner/backdoor processes

pgrep -a -f "xmrig\|stmept\|minerd\|cryptonight"
netstat -tlnp 2>/dev/null | grep -E ":3333|:4444|:5555|:8888"

Phase 5: Auto-Fix (Apply Hardening)

When confirmed by user, apply these fixes:

Fix 1: Create storage/.htaccess (SAFE - zero breaking changes)

<IfModule mod_php.c>
    php_flag engine off
</IfModule>
<IfModule mod_php7.c>
    php_flag engine off
</IfModule>
<IfModule mod_php8.c>
    php_flag engine off
</IfModule>
<FilesMatch "\.(?:php|phtml|phar|php[0-9]|inc)$">
    Order Deny,Allow
    Deny from all
</FilesMatch>

Fix 2: Harden public/.htaccess (SAFE - zero breaking changes)

Add before existing rewrite rules:

# Block PHP execution in storage/upload directories
RewriteRule ^storage/.*\.php$ - [F,L]
RewriteRule ^uploads/.*\.php$ - [F,L]
RewriteRule ^images/.*\.php$ - [F,L]

Add after </IfModule>:

# Security Headers
<IfModule mod_headers.c>
    Header set X-Content-Type-Options "nosniff"
    Header set X-Frame-Options "SAMEORIGIN"
    Header set X-XSS-Protection "1; mode=block"
</IfModule>

# Block access to sensitive files
<FilesMatch "\.(env|log|sql|bak|conf|ini)$">
    Order Deny,Allow
    Deny from all
</FilesMatch>

Fix 3: Generate nginx security config (SAFE - needs manual include)

Create nginx-security.conf:

location ~ ^/storage/.*\.php$ { deny all; return 403; }
location ~ /\.(env|git|htaccess|htpasswd) { deny all; return 404; }
location ~* \.(sql|bak|log|conf|ini)$ { deny all; return 404; }
add_header X-Content-Type-Options "nosniff" always;
add_header X-Frame-Options "SAMEORIGIN" always;
add_header X-XSS-Protection "1; mode=block" always;

Fix 4: Vulnerable package mitigation (CHECK for breaking changes)

For packages with breaking changes (e.g., intervention/image v2 → v3):

  1. Copy package from vendor/ to packages/<vendor>/<package>/
  2. Add path repository to composer.json:
{
    "repositories": [{
        "type": "path",
        "url": "packages/<vendor>/<package>",
        "options": { "symlink": true }
    }]
}
  1. This allows patching the vulnerability locally without breaking API changes

Report Format

Present findings as:

## Security Audit Report - [Project Name]

### CRITICAL
- [Findings that need immediate action]

### HIGH
- [Findings that should be fixed soon]

### MEDIUM
- [Findings to address in next sprint]

### LOW
- [Recommendations for improvement]

### AUTO-FIXED
- [List of hardening applied]

### ACTION ITEMS
1. [ ] Rotate all credentials in .env
2. [ ] Kill any running miner processes
3. [ ] Review sudoers configuration
4. [ ] Deploy .htaccess changes to production
5. [ ] Monitor access logs for 48 hours

Common Attack Chain Reference

Based on real incidents, this is the typical attack flow:

1. ENTRY: Path traversal via vulnerable package (e.g., intervention/image v2)
   → Upload PHP webshell to storage/public directory

2. EXECUTION: Web server executes uploaded .php file
   → Because no .htaccess blocks PHP in storage dirs

3. DATA THEFT: Webshell reads .env file
   → Gets DB credentials, API keys (SendGrid, AWS, Stripe, etc.)

4. LATERAL MOVEMENT: Read .env from other projects on same server
   → Shared server = all projects compromised

5. PERSISTENCE: Install crypto miner + cron job
   → guard.sh / wp-admin.php pattern: download xmrig, auto-restart

6. ABUSE: Use stolen API keys
   → Send phishing emails via SendGrid
   → Access cloud resources via AWS keys

Important Notes

  • NEVER output actual credential values - only indicate what types exist
  • Always ask before applying fixes that could have breaking changes
  • Storage.htaccess and public.htaccess hardening are ALWAYS safe to apply
  • When in doubt, report and let the user decide
  • This skill is designed for defensive security only

平台分布

Codex

38.48%
按下载量换算36

Claude

29.7%
按下载量换算28

Cursor

18.24%
按下载量换算17

Gemini CLI

9.06%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills