Sentinai: Zero-Trust AI Security Guardian

OVERVIEW Sentinai is a proactive, zero-trust AI security guardian engineered to protect codebase repositories and database infrastructure from real-time developer errors and autonomous AI agent threats. THE PROBLEM As software teams increasingly adopt autonomous AI coding agents, the speed of development is soaring. However, traditional security tools (SAST/DAST) are entirely static and reactive. They only scan code at compile or build times, leaving systems completely vulnerable to accidental or malicious runtime CLI commands, database drops, or configuration compromises during execution. THE SOLUTION Sentinai changes security from a build-time afterthought to active, real-time runtime guardrails. Powered by Google Gemini, the platform delivers three layers of deterministic protection: 1. Command Jail: Intercepts terminal commands (such as destructive docker exec or psql database drops) before they execute, sandbox-evaluating them for harmful intent and blocking execution instantly. 2. File Sentinel: Continuous real-time OS-level file monitoring that protects critical system environment configurations, database schemas, and Alembic migrations. 3. PR Diff Guard: An automated virtual security engineer that audits incoming git pull request diffs for hidden backdoors or bypassed authentication middleware. With a fully reactive Nuxt 4 frontend and an asynchronous FastAPI backend securely deployed and synchronized, Sentinai keeps your infrastructure completely safe.