It has a few rough edges but it works and it's free (github.com)

ExoArmur has introduced a new governance tool designed to enhance accountability in AI operations. This system generates replay-verifiable proof bundles for AI agent actions, ensuring that each decision is policy-gated, auditable, replayable, and approvable. Currently ready for technical evaluation, ExoArmur acts as an intermediary layer between AI decision-making and execution, enforcing strict compliance with defined actions. Its functionality includes producing cryptographic audit trails tied to the original intention of actions and guaranteeing deterministic execution that can be reconstructed identically across multiple runs. This innovation is significant for the AI/ML community as it enhances trust and governance in AI systems, addressing critical concerns around transparency and accountability. The ExoArmur architecture allows for clear visibility into decision-making processes and the potential for human oversight, which can mitigate risks associated with autonomous AI operations. With features like three-run stability verification, deterministic replay, and a robust error-correction system, it positions itself as an essential tool for organizations looking to implement safer and more reliable AI frameworks. However, it comes with limitations, such as not providing Byzantine Fault Tolerance and currently lacking production certification, indicating that further development and testing are required before full deployment.