Halo provides open-source, tamper‑evident runtime evidence for AI agents

Halo is an open‑source project that records runtime evidence for AI agents. It focuses on creating tamper‑evident logs of an agent’s actions. The repository is hosted on GitHub for community access. Halo aims to improve transparency and auditability of AI behavior. It captures execution traces and metadata during inference. The evidence can be verified to detect alterations. Developers can integrate Halo into AI pipelines to track decisions. The project includes documentation and examples for usage.