security-audit（安全审计）

Security Audit Skill

Security audit patterns (OWASP Top 10, LLM Top 10 2025, CWE Top 25 2025, CVSS v4.0), cloud/IaC checks, GitHub security. 80+ PHP/TYPO3 checkpoints (v14.3 LTS in typo3-security.md).

Expertise Areas

• Vulnerabilities: XXE, SQLi, XSS, CSRF, command injection, path traversal, file upload, deserialization, SSRF, SSTI, JWT, type juggling
• Standards: OWASP Top 10 / API / LLM (2025), CWE Top 25, CVSS v3.1/v4.0, OWASP ASVS
• Cloud & IaC: AWS, Azure, GCP; Terraform, Kubernetes, Docker, Helm
• API & Frontend: REST/GraphQL authZ, rate limits, mass assignment, CSP, DOM-XSS
• AI Agents: SKILL.md/AGENTS.md/CLAUDE.md/mcp.json/hooks.json audit; prompt injection; excessive agency

Reference Files (in references/, .md implied)

• Core: owasp-top10, cwe-top25, xxe-prevention, cvss-scoring, api-key-encryption
• Prevention: deserialization-prevention, path-traversal-prevention, file-upload-security, input-validation, error-message-sanitization
• Architecture: authentication-patterns, security-headers, security-logging, cryptography-guide
• Language features (*-security-features): php, python, javascript-typescript, nodejs, java, csharp, go, rust, ruby
• Frameworks (*-security): typo3, typo3-fluid, typo3-typoscript, symfony, laravel, django, flask, fastapi, spring, dotnet, blazor, rails, gin, react, vue, angular, nextjs, nuxt, express, nestjs
• Mobile: android-sdk-security, ios-sdk-security
• Cloud & IaC: aws-security, azure-security, gcp-security, iac-security
• API & Frontend: api-security, frontend-security
• AI Agent: llm-security (OWASP LLM Top 10 2025)
• Shared: framework-security
• Threats: modern-attacks, cve-patterns, cve-database
• DevSecOps: ci-security-pipeline, supply-chain-security, automated-scanning, gha-security
• Incident: supply-chain-incident-response

Quick Patterns

XML parsing (prevent XXE):

$doc->loadXML($input, LIBXML_NONET);

SQL (prevent injection):

$stmt = $pdo->prepare('SELECT * FROM users WHERE id = ?');
$stmt->execute([$id]);

Output (prevent XSS):

echo htmlspecialchars($input, ENT_QUOTES | ENT_HTML5, 'UTF-8');

API keys, passwords, randomness:

$n = random_bytes(SODIUM_CRYPTO_SECRETBOX_NONCEBYTES);
$enc = 'enc:' . base64_encode($n . sodium_crypto_secretbox($apiKey, $n, $key));
password_hash($pw, PASSWORD_ARGON2ID);
bin2hex(random_bytes(32));   // never mt_rand/rand

Automated scanners: references/automated-scanning.md.

Security Checklist

• semgrep/opengrep, trivy fs --severity HIGH,CRITICAL, gitleaks clean
• bcrypt/Argon2 passwords, CSRF on state changes, TLS 1.2+
• Server-side input validation; parameterized SQL; XML entities off
• Output encoding + CSP; no unserialize() on user input
• API keys encrypted; exception messages sanitized
• Secrets out of VCS; audit logging on
• Uploads validated, renamed, outside web root
• Headers HSTS + X-Content-Type-Options; dependencies scanned

GitHub Actions Security

• NEVER interpolate ${{inputs.*}} / ${{github.event.*}} in run: — use env:
• Dependency triage: upgrade > override > dismiss. Full patterns: references/gha-security.md.

Verification

./scripts/security-audit-dispatcher.sh /path/to/project  # auto-detect stack
./scripts/security-audit.sh /path/to/project             # PHP-only
./scripts/github-security-audit.sh owner/repo            # GH repo

Dispatcher detects the stack from indicator files and runs matching scripts/scanners/*.sh (17 ecosystems; see references/ index).

Contributing: https://github.com/netresearch/security-audit-skill