claude-red is a curated library of offensive security skills designed for the Claude AI assistant that enables it to function as a context-aware red team operator.
The library addresses the need to equip Claude with expert-level offensive security knowledge across multiple attack surfaces without consuming context tokens unnecessarily. It works by packaging attack methodologies into structured SKILL.md files that prime Claude with techniques, tooling, edge cases, and escalation paths for specific domains like SQL injection, shellcode development, EDR evasion, and ADCS abuse. Skills load on demand based on conversational triggers, so you only pay context costs for the expertise you actively use.
The tool suits authorized red team engagements, bug bounty triage, security research, CTF preparation, operator training, and systematic attack surface exploration. It integrates with Claude's Skills system, which auto-loads matching skills based on conversational context, though skills can also be manually pasted into project system prompts or conversations. The approach treats each attack surface as a discrete, loadable domain of expertise rather than requiring all offensive knowledge to be present simultaneously.
The project maintains an active skill library organized by category, with a documented roadmap for expanding coverage and a contribution framework for adding new skills. Development activity shows ongoing curation and expansion of the skill set, with structured processes for integrating new offensive security methodologies into the library.
