security-agent-skills / rules

# Cursor Project Rules — Security Agent Skills

## Overview

This repository provides a curated library of security testing skills
for AI agents. Skills are organized by domain under `skills/`.

## Skill Discovery

When a security-related task is detected, Cursor should:

1. Check `skills/<domain>/<skill-name>/SKILL.md` for relevant guidance.
2. Follow any references, scripts, or validation steps linked from the skill.
3. Prefer repo-local skill documentation over generic heuristics.

## Domains

- `skills/recon/` — OSINT, reconnaissance, subdomain discovery
- `skills/web-pentest/` — web application penetration testing
- `skills/network-pentest/` — network-level penetration testing
- `skills/exploit-dev/` — exploit development and vulnerability research
- `skills/reverse-engineering/` — binary analysis, decompilation, RE
- `skills/ctf/` — capture-the-flag methodologies and toolkits
- `skills/post-exploitation/` — post-exploitation operations
- `skills/cloud-security/` — cloud security assessment
- `skills/hardware-iot/` — hardware hacking and IoT security
- `skills/security-orchestrator/` — cross-domain attack chain coordination

## Installation

Run `./install.sh --agent cursor` to install skills into `.cursor/skills/`.
Use `--copy` for copy mode (default is symlink).

## Catalog

See `catalog.yaml` for the machine-readable index of all skills,
domains, and metadata.