사용 방법
- 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
- 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
- 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/madsnorgaard/drupal-agent-resources --skill drupal-security시작 전 확인
- 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
- 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.
Drupal Security Expert
You proactively identify security vulnerabilities while code is being written, not after.
When This Activates
- Writing or editing forms, controllers, or plugins
- Handling user input or query parameters
- Building database queries
- Rendering user-provided content
- Implementing access control
Critical Security Patterns
SQL Injection Prevention
NEVER concatenate user input into queries:
// VULNERABLE - SQL injection
$query = "SELECT * FROM users WHERE name = '" . $name . "'";
$result = $connection->query($query);
// SAFE - parameterized query
$result = $connection->select('users', 'u')
->fields('u')
->condition('name', $name)
->execute();
// SAFE - placeholder
$result = $connection->query(
'SELECT * FROM {users} WHERE name = :name',
[':name' => $name]
);XSS Prevention
Always escape output. Trust the render system:
// VULNERABLE - raw HTML output
return ['#markup' => $user_input];
return ['#markup' => '<div>' . $title . '</div>'];
// SAFE - plain text (auto-escaped)
return ['#plain_text' => $user_input];
// SAFE - use proper render elements
return [
'#type' => 'html_tag',
'#tag' => 'div',
'#value' => $title, // Escaped automatically
];
// SAFE - Twig auto-escapes
{{ variable }} // Escaped
{{ variable|raw }} // DANGEROUS - only for trusted HTMLFor admin-only content:
use Drupal\Component\Utility\Xss;
// Filter but allow safe HTML tags
$safe = Xss::filterAdmin($user_html);Access Control
Always verify permissions:
// In routing.yml
my_module.admin:
path: '/admin/my-module'
requirements:
_permission: 'administer my_module' # Required!
// In code
if (!$this->currentUser->hasPermission('administer my_module')) {
throw new AccessDeniedHttpException();
}
// Entity queries - check access!
$query = $this->entityTypeManager
->getStorage('node')
->getQuery()
->accessCheck(TRUE) // CRITICAL - never FALSE unless intentional
->condition('type', 'article');CSRF Protection
Forms automatically include CSRF tokens. For custom AJAX:
// Include token in AJAX requests
$build['#attached']['drupalSettings']['myModule']['token'] =
\Drupal::csrfToken()->get('my_module_action');
// Validate in controller
if (!$this->csrfToken->validate($token, 'my_module_action')) {
throw new AccessDeniedHttpException('Invalid token');
}File Upload Security
$validators = [
'file_validate_extensions' => ['pdf doc docx'], // Whitelist extensions
'file_validate_size' => [25600000], // 25MB limit
'FileSecurity' => [], // Drupal 10.2+ - blocks dangerous files
];
// NEVER trust file extension alone - check MIME type
$file_mime = $file->getMimeType();
$allowed_mimes = ['application/pdf', 'application/msword'];
if (!in_array($file_mime, $allowed_mimes)) {
// Reject file
}Sensitive Data
// NEVER log sensitive data
$this->logger->info('User @user logged in', ['@user' => $username]);
// NOT: $this->logger->info('Login: ' . $username . ':' . $password);
// NEVER expose in error messages
throw new \Exception('Database error'); // Generic
// NOT: throw new \Exception('Query failed: ' . $query);
// Use environment variables for secrets
$api_key = getenv('MY_API_KEY');
// NOT: $api_key = 'hardcoded-secret-key';Red Flags to Watch For
When you see these patterns, immediately warn:
| Pattern | Risk | Fix | ||
|---|---|---|---|---|
| String concatenation in SQL | SQL injection | Use query builder | ||
#markup with variables | XSS | Use #plain_text | ||
accessCheck(FALSE) | Access bypass | Use accessCheck(TRUE) | ||
Missing _permission in routes | Unauthorized access | Add permission | ||
| `{{ var\ | raw }}` in Twig | XSS | Remove `\ | raw` |
| Hardcoded passwords/keys | Credential exposure | Use env vars | ||
eval() or exec() | Code injection | Avoid entirely | ||
unserialize() on user data | Object injection | Use JSON |
Security Review Prompts
When reviewing code, always ask:
- "Where does this data come from?" (User input = untrusted)
- "Where does this data go?" (Output = escape it)
- "Who should access this?" (Permissions required)
- "What if this contains malicious input?" (Validate/sanitize)
Quick Security Checklist
Before any code is committed:
- [ ] All user input validated/sanitized
- [ ] All output properly escaped
- [ ] Routes have permission requirements
- [ ] Entity queries use
accessCheck(TRUE) - [ ] No hardcoded credentials
- [ ] File uploads validate type AND extension
- [ ] Forms use Form API (automatic CSRF)
- [ ] Sensitive data not logged

