---

name: secure-coding description: Provides guidance on secure coding practices including OWASP Top 10 2025, CWE Top 25, input validation, output encoding, and language-specific security patterns. Use when reviewing code for security vulnerabilities, implementing security controls, or learning secure development practices. allowed-tools: Read, Glob, Grep, Task

Secure Coding

Comprehensive guidance for writing secure code, covering OWASP Top 10 2025, CWE Top 25, and language-specific security patterns.

When to Use This Skill

Use this skill when:

• Reviewing code for security vulnerabilities
• Implementing input validation or output encoding
• Learning about common security weaknesses (OWASP, CWE)
• Fixing identified security issues
• Writing security-sensitive code (authentication, authorization, data handling)
• Conducting security code reviews

OWASP Top 10 2025 Quick Reference

Rank	Vulnerability	Key Mitigation
A01	Broken Access Control	Server-side access checks, deny by default, CORS restrictions
A02	Security Misconfiguration	Hardened configs, remove defaults, disable unnecessary features
A03	Software Supply Chain Failures	SCA, SBOM, verify dependencies, integrity checks
A04	Cryptographic Failures	Strong encryption (AES-256), TLS 1.2+, no deprecated algorithms
A05	Injection	Parameterized queries, input validation, context-aware encoding
A06	Insecure Design	Threat modeling, secure design patterns, defense in depth
A07	Authentication Failures	MFA, strong passwords, secure session management
A08	Data Integrity Failures	Digital signatures, integrity verification, secure CI/CD
A09	Logging & Alerting Failures	Centralized logging, anomaly detection, audit trails
A10	Mishandling Exceptions	Fail securely, generic error messages, complete exception handling

For detailed mitigations: See OWASP Top 10 2025 Reference

Core Secure Coding Principles

1. Input Validation

Never trust user input. Validate all inputs on the server side.

using System.Text.RegularExpressions;

// Good: Server-side validation with allowlist
public static partial class InputValidation
{
    [GeneratedRegex(@"^[a-zA-Z0-9_]{3,20}$")]
    private static partial Regex UsernamePattern();

    /// <summary>
    /// Validate username against allowlist pattern.
    /// </summary>
    public static bool ValidateUsername(string username) =>
        !string.IsNullOrEmpty(username) && UsernamePattern().IsMatch(username);
}

// Bad: No validation
public string ProcessUsername(string username) => username;  // Dangerous!

Validation strategies:

• Allowlist validation: Define what IS allowed (preferred)
• Blocklist validation: Define what is NOT allowed (less secure)
• Type checking: Ensure correct data types
• Range checking: Verify values within expected bounds
• Length limits: Prevent buffer overflows and DoS

2. Output Encoding

Encode output based on context to prevent injection attacks.

Context	Encoding Method	Example
HTML body	HTML entity encoding	<script>
HTML attributes	Attribute encoding	' for '
JavaScript	JavaScript encoding	\x3Cscript\x3E
URL parameters	URL encoding	%3Cscript%3E
CSS	CSS encoding	\3C script\3E
SQL	Parameterized queries	Use prepared statements

3. Parameterized Queries (Injection Prevention)

Always use parameterized queries for database operations.

// Good: Parameterized query with SqlCommand
using var cmd = new SqlCommand(
    "SELECT * FROM Users WHERE Username = @username AND Status = @status",
    connection);
cmd.Parameters.AddWithValue("@username", username);
cmd.Parameters.AddWithValue("@status", status);

// Good: Parameterized query with Dapper
var users = await connection.QueryAsync<User>(
    "SELECT * FROM Users WHERE Username = @Username AND Status = @Status",
    new { Username = username, Status = status });

// Bad: String interpolation (SQL Injection vulnerable)
var query = $"SELECT * FROM Users WHERE Username = '{username}'";  // VULNERABLE

4. Authentication Security

• Use strong password hashing: Argon2id, bcrypt, scrypt (see cryptography skill)
• Implement MFA: Time-based OTP, hardware keys, passkeys
• Secure session management: HttpOnly cookies, secure flag, short expiration
• Account lockout: Prevent brute force attacks
• Credential storage: Never store plaintext passwords

5. Authorization Security

• Deny by default: Require explicit permission grants
• Server-side checks: Never rely on client-side authorization
• Verify object ownership: Check user can access requested resource
• Use indirect references: Map internal IDs to user-specific references
• Implement RBAC/ABAC: Use structured access control models

6. Error Handling

// Good: Generic error message to user, detailed logging
public async Task<IActionResult> ProcessData([FromBody] DataRequest request)
{
    try
    {
        await _dataService.ProcessSensitiveDataAsync(request.Data);
        return Ok();
    }
    catch (DbException ex)
    {
        _logger.LogError(ex, "Database error processing request for user {UserId}", User.GetUserId());
        return StatusCode(500, new { error = "An error occurred" });
    }
}

// Bad: Exposing internal details
catch (DbException ex)
{
    return StatusCode(500, new { error = ex.Message });  // VULNERABLE - exposes internals
}

Error handling rules:

• Return generic error messages to users
• Log detailed errors server-side
• Never expose stack traces, database errors, or internal paths
• Fail securely (deny access on error)

Language-Specific Patterns

JavaScript/TypeScript

// XSS Prevention - use textContent, not innerHTML
element.textContent = userInput;  // Safe
element.innerHTML = userInput;    // VULNERABLE

// Use DOMPurify for HTML that must be rendered
import DOMPurify from 'dompurify';
element.innerHTML = DOMPurify.sanitize(userInput);

// Avoid eval() and Function()
eval(userInput);           // VULNERABLE
new Function(userInput)(); // VULNERABLE

// Use strict mode
'use strict';

C# / .NET

// Safe process execution - use argument list, avoid shell
using System.Diagnostics;

public static async Task<string> SafeExecuteAsync(string command, params string[] args)
{
    using var process = new Process
    {
        StartInfo = new ProcessStartInfo
        {
            FileName = command,
            RedirectStandardOutput = true,
            RedirectStandardError = true,
            UseShellExecute = false,  // Safe: no shell interpretation
            CreateNoWindow = true
        }
    };

    foreach (var arg in args)
        process.StartInfo.ArgumentList.Add(arg);  // Safe: no shell escaping needed

    process.Start();
    var output = await process.StandardOutput.ReadToEndAsync();
    await process.WaitForExitAsync();
    return output;
}

// Bad: Shell injection vulnerable
Process.Start("cmd", $"/c dir {userInput}");  // VULNERABLE

// Safe file operations - prevent path traversal
public static class SafeFileAccess
{
    /// <summary>
    /// Safely read a file, preventing path traversal attacks.
    /// </summary>
    public static string SafeReadFile(string baseDir, string filename)
    {
        var basePath = Path.GetFullPath(baseDir);
        var filePath = Path.GetFullPath(Path.Combine(baseDir, filename));

        if (!filePath.StartsWith(basePath, StringComparison.OrdinalIgnoreCase))
            throw new UnauthorizedAccessException("Path traversal detected");

        return File.ReadAllText(filePath);
    }
}

// Use parameterized queries with Entity Framework
var users = context.Users
    .Where(u => u.Username == username)  // Safe - parameterized
    .ToList();

// Avoid raw SQL when possible, use parameters if needed
var users = context.Users
    .FromSqlRaw("SELECT * FROM Users WHERE Username = {0}", username)
    .ToList();

// Anti-forgery tokens for CSRF protection
[ValidateAntiForgeryToken]
public IActionResult UpdateProfile(ProfileModel model)
{
    // Process update
}

// Input validation with data annotations
public sealed class UserInput
{
    [Required]
    [StringLength(100, MinimumLength = 3)]
    [RegularExpression(@"^[a-zA-Z0-9_]+$")]
    public required string Username { get; init; }
}

Security Code Review Checklist

Input Handling

• All inputs validated on server side
• Allowlist validation used where possible
• Length limits enforced
• Type checking performed
• File uploads validated (type, size, content)

Output Encoding

• Context-appropriate encoding applied
• No raw user input in HTML/JS/SQL
• Content-Type headers set correctly
• X-Content-Type-Options: nosniff

Authentication

• Strong password hashing (Argon2id/bcrypt)
• Session tokens are random and unpredictable
• Session invalidation on logout
• Account lockout after failed attempts
• Credentials transmitted over HTTPS only

Authorization

• Access control on every request
• Deny by default policy
• Object-level authorization checks
• No direct object references exposed

Data Protection

• Sensitive data encrypted at rest
• TLS 1.2+ for data in transit
• No sensitive data in URLs or logs
• Proper key management

Error Handling

• Generic error messages to users
• Detailed errors logged securely
• No stack traces exposed
• Fail securely (deny on error)

Quick Decision Tree

What security concern are you addressing?

1. SQL/NoSQL injection → Use parameterized queries, ORMs
2. XSS (Cross-Site Scripting) → Context-aware output encoding, CSP
3. CSRF → Anti-forgery tokens, SameSite cookies
4. Authentication → See authentication-patterns skill
5. Authorization → See authorization-models skill
6. Cryptography → See cryptography skill
7. Secrets/Credentials → See secrets-management skill
8. API Security → See api-security skill

References

• OWASP Top 10 2025 Detailed Reference - Complete mitigations and examples
• CWE Top 25 Reference - Most dangerous software weaknesses
• Language-Specific Patterns - Per-language security guides

Related Skills

Skill	Relationship
authentication-patterns	Auth implementation details (JWT, OAuth, Passkeys)
authorization-models	Access control (RBAC, ABAC)
cryptography	Encryption, hashing, TLS
api-security	API-specific security patterns
secrets-management	Credential and secret handling

Version History

• v1.0.0 (2025-12-26): Initial release with OWASP Top 10 2025, core principles

---

Last Updated: 2025-12-26