TrustGuard AI scans your prompts in CI and blocks jailbreaks in production — no ML-security expertise required.