offensive-ai-security (Claude-Red)
AI-pentest skill for red-teaming LLM apps: prompt injection, jailbreaks, and agentic exploitation.
🛡️ AgentReady threat assessment
MAESTRO 7-layer threat model + OWASP AIVSS risk score for offensive-ai-security (Claude-Red), derived from its capabilities.
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology — every score is re-derived by the same automated method as an agent's public evidence changes.
Overview
An offensive skill focused on penetration testing of AI/LLM systems — prompt injection, jailbreaks, model and agent abuse — derived from the author's offensive-checklist ai.md. Surface: guides adversarial prompt crafting and exploitation of LLM-backed applications.
Key features and capabilities
- LLM prompt-injection and jailbreak testing
- Agentic/AI-app exploitation methodology
- Based on a maintained offensive checklist
Use cases
- Red-team an LLM-powered application
- Test an AI agent for injection and abuse