Hack the model: Build AI security skills with the GitHub Secure Code Game

GitHub announces season three of the free Secure Code Game, a code-first training that immerses developers in AI security through six escalating challenges covering robust system prompts, input filtering, output validation, and LLM self-verification; players first exploit LLM prompt vulnerabilities to reveal secrets, then harden code and system messages while maintaining functionality, with quick setup via Codespaces, access to GitHub Models, testimonials from participants, and an open invitation to contribute new challenges.