hack-skills

Why we included this project

Security teams running AI agents on real engagements will find this a practical shortcut: instead of a generic assistant that guesses at methodology, the agent gets a structured knowledge base that knows how to approach a target. The repo packs 101 deep-topic skills across 14 security domains, including web and API security, OS privilege escalation, Active Directory attacks, binary exploitation, and AI/LLM security, into one installable skill set. The layering is what makes it usable: a master entry routes the agent, category entries pick an attack surface, and deep-topic skills load on demand, so the agent only pulls in what the current phase needs. Skills are plain SKILL.md files, easy to audit and diff before you trust them in a live engagement. There's also a static searchable web UI and an encrypted offline ZIP for air-gapped targets, a nice touch when the target network is off the internet.