security（安全）

Security Skill

Security best practices and automated security testing for all projects.

Core Principle

Security is not optional. Every project must pass security checks before merge. Assume all input is malicious, all secrets will leak if committed, and all dependencies have vulnerabilities.

Required Security Setup

1. Gitignore (Non-Negotiable)

Every project must have these in .gitignore:

# Environment files - NEVER commit
.env
.env.*
!.env.example

# Secrets
*.pem
*.key
*.p12
*.pfx
credentials.json
secrets.json
*-credentials.json
service-account*.json

# IDE and OS
.idea/
.vscode/settings.json
.DS_Store
Thumbs.db

# Dependencies
node_modules/
__pycache__/
*.pyc
.venv/
venv/

# Build outputs
dist/
build/
*.egg-info/

# Logs that might contain sensitive data
*.log
logs/

2. Environment Variables

Create .env.example with all required vars (no values):

# .env.example - Copy to .env and fill in values

# Server-side only (NEVER prefix with VITE_ or NEXT_PUBLIC_)
DATABASE_URL=
ANTHROPIC_API_KEY=
SUPABASE_SERVICE_ROLE_KEY=

# Client-side safe (public, non-sensitive)
VITE_SUPABASE_URL=
VITE_SUPABASE_ANON_KEY=

Frontend Environment Variables (Critical!)

NEVER put secrets in client-exposed env vars:

// WRONG - Secret exposed to browser bundle!
const apiKey = import.meta.env.VITE_ANTHROPIC_API_KEY;

// CORRECT - Only public values client-side
const supabaseUrl = import.meta.env.VITE_SUPABASE_URL;

// CORRECT - Secrets stay server-side only
// In API route or server function:
const apiKey = process.env.ANTHROPIC_API_KEY;

Vercel Environment Variables:

• In Vercel dashboard, secrets without VITE_ prefix are server-only
• Only VITE_* vars are bundled into client code
• Always verify in browser devtools → Sources → your bundle that secrets aren't exposed

Validate environment at startup:

// config/env.ts
import { z } from 'zod';

const envSchema = z.object({
  DATABASE_URL: z.string().url(),
  ANTHROPIC_API_KEY: z.string().min(1),
  NODE_ENV: z.enum(['development', 'production', 'test']),
});

export const env = envSchema.parse(process.env);

# config/env.py
from pydantic_settings import BaseSettings

class Settings(BaseSettings):
    database_url: str
    anthropic_api_key: str
    environment: str = "development"

    class Config:
        env_file = ".env"

settings = Settings()

Security Tests

Pre-Commit Security Checks

Add to pre-commit hooks:

For all projects:

# .pre-commit-config.yaml (add to existing)
repos:
  # Detect secrets
  - repo: https://github.com/Yelp/detect-secrets
    rev: v1.4.0
    hooks:
      - id: detect-secrets
        args: ['--baseline', '.secrets.baseline']

  # Check for security issues in dependencies
  - repo: local
    hooks:
      - id: security-check
        name: security-check
        entry: ./scripts/security-check.sh
        language: script
        pass_filenames: false

TypeScript/JavaScript:

// package.json scripts
{
  "scripts": {
    "security:audit": "npm audit --audit-level=high",
    "security:secrets": "npx secretlint '**/*'",
    "security:deps": "npx better-npm-audit audit"
  }
}

Python:

# Add to dev dependencies
pip install safety bandit

# Commands
safety check           # Check dependencies for vulnerabilities
bandit -r src/        # Static security analysis

Security Check Script

Create scripts/security-check.sh:

#!/bin/bash
set -e

echo "Running security checks..."

# Check for secrets in staged files
echo "Checking for secrets..."
if command -v detect-secrets &> /dev/null; then
  detect-secrets scan --baseline .secrets.baseline
fi

# Check .env is not staged
if git diff --cached --name-only | grep -E '^\.env$|^\.env\.' | grep -v '\.example$'; then
  echo "ERROR: .env file is staged for commit!"
  exit 1
fi

# Check for common secret patterns in staged files
STAGED_FILES=$(git diff --cached --name-only --diff-filter=ACM)
if echo "$STAGED_FILES" | xargs grep -l -E '(password|secret|api_key|apikey|token|private_key)\s*[:=]\s*["\047][^"\047]+["\047]' 2>/dev/null; then
  echo "ERROR: Possible secrets found in staged files!"
  exit 1
fi

# Language-specific checks
if [ -f "package.json" ]; then
  echo "Checking npm dependencies..."
  npm audit --audit-level=high || echo "Warning: npm audit found issues"
fi

if [ -f "pyproject.toml" ] || [ -f "requirements.txt" ]; then
  echo "Checking Python dependencies..."
  if command -v safety &> /dev/null; then
    safety check || echo "Warning: safety found issues"
  fi
fi

echo "Security checks passed!"

chmod +x scripts/security-check.sh

GitHub Actions Security Workflow

Create .github/workflows/security.yml:

name: Security

on:
  push:
    branches: [main]
  pull_request:
    branches: [main]
  schedule:
    # Run weekly on Monday at 9am UTC
    - cron: '0 9 * * 1'

jobs:
  secrets-scan:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Detect secrets
        uses: trufflesecurity/trufflehog@main
        with:
          path: ./
          base: ${{ github.event.pull_request.base.sha }}
          head: ${{ github.event.pull_request.head.sha }}

  dependency-audit:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      # Node.js projects
      - name: Setup Node
        if: hashFiles('package.json') != ''
        uses: actions/setup-node@v4
        with:
          node-version: '20'

      - name: Install dependencies
        if: hashFiles('package.json') != ''
        run: npm ci

      - name: NPM Audit
        if: hashFiles('package.json') != ''
        run: npm audit --audit-level=high

      # Python projects
      - name: Setup Python
        if: hashFiles('pyproject.toml') != '' || hashFiles('requirements.txt') != ''
        uses: actions/setup-python@v5
        with:
          python-version: '3.12'

      - name: Install safety
        if: hashFiles('pyproject.toml') != '' || hashFiles('requirements.txt') != ''
        run: pip install safety

      - name: Safety check
        if: hashFiles('pyproject.toml') != '' || hashFiles('requirements.txt') != ''
        run: safety check

  codeql:
    runs-on: ubuntu-latest
    permissions:
      security-events: write
    steps:
      - uses: actions/checkout@v4

      - name: Initialize CodeQL
        uses: github/codeql-action/init@v3
        with:
          languages: ${{ hashFiles('package.json') != '' && 'javascript-typescript' || 'python' }}

      - name: Autobuild
        uses: github/codeql-action/autobuild@v3

      - name: Perform CodeQL Analysis
        uses: github/codeql-action/analyze@v3

Input Validation (OWASP Top 10)

1. SQL Injection Prevention

Never use string concatenation:

// BAD - SQL injection vulnerable
const user = await db.query(`SELECT * FROM users WHERE id = ${userId}`);

// GOOD - Parameterized query
const user = await db.query('SELECT * FROM users WHERE id = $1', [userId]);

// GOOD - Using ORM (Kysely, Prisma, Drizzle)
const user = await db.selectFrom('users').where('id', '=', userId).execute();

# BAD - SQL injection vulnerable
cursor.execute(f"SELECT * FROM users WHERE id = {user_id}")

# GOOD - Parameterized query
cursor.execute("SELECT * FROM users WHERE id = %s", (user_id,))

# GOOD - Using ORM (SQLAlchemy)
user = session.query(User).filter(User.id == user_id).first()

2. XSS Prevention

// Always sanitize user input before rendering
import DOMPurify from 'dompurify';

// BAD - XSS vulnerable
element.innerHTML = userInput;

// GOOD - Sanitized
element.innerHTML = DOMPurify.sanitize(userInput);

// BEST - Use framework's built-in escaping (React does this by default)
return <div>{userInput}</div>;  // Safe in React

// DANGER - Bypasses React's protection
return <div dangerouslySetInnerHTML={{ __html: userInput }} />;  // Avoid!

3. Input Validation at Boundaries

// Validate ALL external input with Zod
import { z } from 'zod';

const CreateUserSchema = z.object({
  email: z.string().email().max(255),
  name: z.string().min(1).max(100).regex(/^[a-zA-Z\s]+$/),
  age: z.number().int().min(0).max(150),
});

// In route handler
app.post('/users', async (req, res) => {
  const result = CreateUserSchema.safeParse(req.body);
  if (!result.success) {
    return res.status(400).json({ error: result.error });
  }
  // result.data is now typed and validated
});

4. Path Traversal Prevention

import path from 'path';

// BAD - Path traversal vulnerable
const filePath = `./uploads/${req.params.filename}`;

// GOOD - Validate and sanitize path
const filename = path.basename(req.params.filename);  // Strips ../
const filePath = path.join('./uploads', filename);

// Verify it's still within allowed directory
if (!filePath.startsWith(path.resolve('./uploads'))) {
  throw new Error('Invalid path');
}

Authentication & Authorization

JWT Best Practices

import jwt from 'jsonwebtoken';

// Token generation
function generateToken(userId: string): string {
  return jwt.sign(
    { sub: userId },
    process.env.JWT_SECRET!,
    {
      expiresIn: '15m',      // Short-lived access tokens
      algorithm: 'HS256',
    }
  );
}

// Token verification
function verifyToken(token: string): { sub: string } {
  return jwt.verify(token, process.env.JWT_SECRET!, {
    algorithms: ['HS256'],   // Explicitly specify allowed algorithms
  }) as { sub: string };
}

Password Hashing

import bcrypt from 'bcrypt';

const SALT_ROUNDS = 12;  // Minimum 10, recommended 12+

async function hashPassword(password: string): Promise<string> {
  return bcrypt.hash(password, SALT_ROUNDS);
}

async function verifyPassword(password: string, hash: string): Promise<boolean> {
  return bcrypt.compare(password, hash);
}

from passlib.context import CryptContext

pwd_context = CryptContext(schemes=["bcrypt"], deprecated="auto")

def hash_password(password: str) -> str:
    return pwd_context.hash(password)

def verify_password(password: str, hashed: str) -> bool:
    return pwd_context.verify(password, hashed)

Rate Limiting

import rateLimit from 'express-rate-limit';

const limiter = rateLimit({
  windowMs: 15 * 60 * 1000,  // 15 minutes
  max: 100,                   // 100 requests per window
  standardHeaders: true,
  legacyHeaders: false,
});

// Apply to auth routes
app.use('/api/auth', rateLimit({
  windowMs: 60 * 1000,  // 1 minute
  max: 5,                // 5 attempts per minute
  message: 'Too many login attempts, please try again later',
}));

Security Headers

import helmet from 'helmet';

app.use(helmet({
  contentSecurityPolicy: {
    directives: {
      defaultSrc: ["'self'"],
      scriptSrc: ["'self'"],
      styleSrc: ["'self'", "'unsafe-inline'"],
      imgSrc: ["'self'", "data:", "https:"],
    },
  },
  hsts: {
    maxAge: 31536000,
    includeSubDomains: true,
  },
}));

Security Testing Checklist

Run before every release:

## Security Checklist

### Secrets & Environment
- [ ] No secrets in code (run detect-secrets)
- [ ] .env files in .gitignore
- [ ] .env.example exists with all required vars
- [ ] Environment validated at startup

### Dependencies
- [ ] npm audit / safety check passes
- [ ] No known vulnerabilities in dependencies
- [ ] Dependencies up to date (Dependabot enabled)

### Input Validation
- [ ] All API inputs validated with schema (Zod/Pydantic)
- [ ] File uploads restricted by type and size
- [ ] Path traversal prevented

### Authentication
- [ ] Passwords hashed with bcrypt (12+ rounds)
- [ ] JWTs use short expiration
- [ ] Rate limiting on auth endpoints
- [ ] Session tokens rotated on login

### Database
- [ ] Parameterized queries only
- [ ] Least privilege database user
- [ ] Connection strings not logged

### Headers & CORS
- [ ] Security headers enabled (helmet)
- [ ] CORS restricted to known origins
- [ ] HTTPS only in production

### Logging
- [ ] No secrets in logs
- [ ] No PII in logs (or properly masked)
- [ ] Failed auth attempts logged

Security Anti-Patterns

• ❌ Secrets in VITE_*, NEXT_PUBLIC_*, or REACT_APP_* env vars (client-exposed!)
• ❌ Secrets in code or config files committed to git
• ❌.env files without.gitignore entry
• ❌ String concatenation for SQL queries
• ❌ dangerouslySetInnerHTML without sanitization
• ❌ eval() or new Function() with user input
• ❌ Passwords stored as plain text or weak hash (MD5, SHA1)
• ❌ JWTs with no expiration or very long expiration
• ❌ No rate limiting on authentication endpoints
• ❌ Logging sensitive data (passwords, tokens, PII)
• ❌ Using * for CORS origins in production
• ❌ Ignoring npm audit / safety check warnings
• ❌ Running as root / admin in production
• ❌ Hardcoded credentials for any environment
• ❌ Disabling SSL/TLS verification