Claude-Red open-sources 70+ offensive security skills for Claude AI
Claude-Red is a curated open-source library of over 70 structured SKILL.md files that add offensive security expertise to Anthropic's Claude AI assistant for authorized red team operations.
Summary
SnailSploit has published Claude-Red, an open-source collection of offensive security skills for the Claude Skills system. The repository contains more than 70 SKILL.md files covering attack surfaces from SQL injection and EDR evasion to wireless exploitation and Kubernetes cluster attacks. Each skill loads on demand via conversational triggers, adding domain-specific methodology without consuming context window space until needed. The project targets authorized red team engagements, bug bounty work, security research, and CTF preparation.
Skill categories and deployment
- 70+ skills across 21 categories including Web Application (16), Wireless (14), Infrastructure & Red Team (7), and Exploit Development (6)
- On-demand loading: Skills activate via conversational triggers rather than persistent context, reducing token costs
- Two deployment paths: Native Claude Skills system auto-discovery, or manual paste into Claude.ai system prompts
- Sparse-checkout support:
git sparse-checkoutallows installing single categories without full repository clone - Structured format: Each SKILL.md primes Claude with techniques, tooling, edge cases, and escalation paths for a specific attack surface
Original reference: github.com →