ai-security-arsenal

AI Security Arsenal is a collection of modular skills designed to automate security testing, penetration testing, and security assessment workflows for Claude Code, Cursor, and other AI coding agents. It covers mobile security (Android and iOS), web application testing (DAST and API security), cloud infrastructure assessment, network penetration testing, static code analysis, supply chain security, and threat modeling. Each skill follows a consistent router-based architecture with lazy-loaded workflows, methodologies, payloads, and schemas that allow agents to load only what they need for each task. Security researchers and penetration testers use it to accelerate manual testing phases, automate repetitive assessment tasks, and generate structured findings that integrate with SARIF converters, Jira, and DefectDojo. The arsenal was last validated in April 2026 against frontier coding agents like Claude Opus 4.x.

Each skill is structured as a router file (SKILL.md) that indexes workflows, methodologies, payloads, and schemas. Load the appropriate skill for your task (e.g., android-pentest, dast-automation, api-security), then use the decision tree to select the relevant workflow. Findings validate against per-skill JSON schemas for integration with external tools.