How does it handle sensitive user data?

The skill includes PII masking patterns using tools like Microsoft Presidio and Langfuse to redact sensitive information before it hits logs or LLM providers.

Does this skill help with OWASP Top 10 compliance?

Yes, it includes specific rules and patterns for preventing SQL injection, broken authentication, SSRF, and other critical OWASP vulnerabilities.

Can I use this for LLM-specific security?

Absolutely. It provides advanced guardrails for prompt injection defense, context separation, and output validation to keep your AI integrations safe.

Is it compatible with modern frontend frameworks?

Yes, it features patterns for Zod validation, JWT management, and OAuth 2.1/PKCE that are standard in modern TypeScript and Python full-stack development.

Security Patterns for Claude Code

Name: Security Patterns for Claude Code
Author: yonatangross

byyonatangross

•

116

•

Security & Testing

Implements production-grade security patterns for authentication, input validation, OWASP compliance, and LLM safety.

This skill provides a comprehensive library of security patterns designed to harden full-stack applications and AI-integrated systems. It covers critical areas such as modern authentication (OAuth 2.1, JWT, Argon2id), robust input validation using Zod and Pydantic, and defense-in-depth architectures. Additionally, it addresses unique AI challenges like prompt injection defense, PII masking for observability, and LLM output guardrails, ensuring that developers can build secure, compliant, and resilient software directly within their Claude Code workflow.

Key Features

01Modern Authentication (OAuth 2.1, PKCE, JWT, Argon2id)

02Automated Security Scanning (Semgrep, Bandit, Gitleaks)

03Input Validation & Sanitization (Zod v4, Pydantic, OWASP Top 10)

04116 GitHub stars

05LLM Safety & Prompt Injection Defense

06PII Detection and Redaction (Microsoft Presidio, Langfuse)

Use Cases

01Hardening an API with multi-layer defense and role-based access control

02Securing LLM-powered applications against prompt injections and data leaks

03Implementing GDPR-compliant PII masking in logging and observability pipelines

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yonatangross/orchestkit security-patterns

For use in Claude.ai and ChatGPT

Download Skill