Codacy's AI Risk Hub is designed to ensure security in AI-driven coding. It addresses the potential vulnerabilities that come with using AI coding tools, which can inadvertently introduce risks into codebases. The hub provides a unified approach to enforce secure AI coding policies across various teams and projects, instantly. It focuses on preventing the deployment of code that might contain dangerous API calls to unapproved large language models and helps avoid issues like SQL injection caused by unsanitized user input. Key features include real-time scanning for AI-specific risks, such as invisible unicode injections and hardcoded secrets, which can lead to significant security breaches. The platform also conducts daily vulnerability scans, ensuring that any new CVE risks are promptly identified. Codacy emphasizes the importance of not only scanning code but also implementing essential merge controls. This helps protect against risky AI contributions and ensures that all projects adhere to defined security policies. The system tracks model usage and enforces guidelines to mitigate data leakage risks. It facilitates software composition analysis to catch insecure dependencies introduced by AI coding agents. With a comprehensive checklist to manage AI risk scores, the Codacy AI Risk Hub aims to streamline the coding process while significantly enhancing security measures.