Defensive security routing skill and reference corpus for application security, identity, infrastructure, privacy, incident response, and AI-agent review. Use when the task is a secure code review, audit, hardening, threat modeling, vulnerability management, prompt injection, MCP security, or incident handling, including French triggers such as audit de sécurité, durcir, revue sécurité, failles, secrets exposés, modélisation des menaces, and réponse à incident. Do not use for code golf, general feature ideation without a security angle, or active offensive exploitation, malware, payloads, intrusion steps, or operator-bypass requests.
2026-07-12