security-reviewer

📁 ai-engineer-agent/ai-engineer-skills 📅 5 days ago
2
总安装量
2
周安装量
#64316
全站排名
安装命令
npx skills add https://github.com/ai-engineer-agent/ai-engineer-skills --skill security-reviewer

Agent 安装分布

trae 2
gemini-cli 2
claude-code 2
codex 2
kiro-cli 2
cursor 2

Skill 文档

Security Review

Perform a comprehensive security review of $ARGUMENTS:

Check for OWASP Top 10

  1. Injection (SQL, NoSQL, OS Command, LDAP)

    • Look for string concatenation in queries
    • Check for parameterized queries / prepared statements
    • Check for ORM misuse (raw queries)
  2. Broken Authentication

    • Check password hashing (bcrypt/scrypt/argon2, not MD5/SHA1)
    • Check session management and token expiry
    • Look for hardcoded credentials
  3. Sensitive Data Exposure

    • Check for secrets in code (API keys, passwords, tokens)
    • Verify encryption at rest and in transit
    • Check logging for PII/sensitive data leaks
  4. XML External Entities (XXE)

    • Check XML parser configuration
    • Verify external entity processing is disabled
  5. Broken Access Control

    • Check authorization on every endpoint
    • Look for IDOR vulnerabilities
    • Verify RBAC/ABAC implementation
  6. Security Misconfiguration

    • Check HTTP headers (CORS, CSP, HSTS)
    • Look for debug mode in production
    • Check default credentials
  7. Cross-Site Scripting (XSS)

    • Check output encoding/escaping
    • Look for dangerouslySetInnerHTML, innerHTML, | safe
    • Verify CSP headers
  8. Insecure Deserialization

    • Check for untrusted deserialization
    • Verify input validation before deserialization
  9. Using Components with Known Vulnerabilities

    • Check for outdated dependencies
    • Run npm audit / pip audit / cargo audit
  10. Insufficient Logging & Monitoring

    • Verify auth events are logged
    • Check for sensitive data in logs

Output Format

Provide findings as:

  • CRITICAL: Exploitable vulnerabilities
  • HIGH: Likely exploitable or high-impact issues
  • MEDIUM: Defense-in-depth improvements
  • LOW: Best practice recommendations

For each finding: describe the issue, show the vulnerable code, explain the risk, and provide the fix.