Add the first continuous-action Unity policy: Crawler 60k safetensors and fixed-batch-10 ctx16/ctx32/ctx64 ONNX files. Over seeds 100..103, ctx32 averaged 1721.88 return and is the recommended deployment variant.
Add the Pyramids 60k safetensors bundle and fixed-batch-16 ctx16/ctx32/ctx64 ONNX policies. Closed-loop validation covered 64 episodes per context over Unity seeds 100..103; ctx16 is the recommended variant at 63/64 reaches and mean return 1.7503.
Replace the legacy ONNX-only policy set with reproducible bundle-format-v2 DungeonEscape and SoccerTwos policies. DungeonEscape KV32 validation: 69/108 ego-agent and 34/36 cooperative-group successes. SoccerTwos: 33/80 wins against the release-23 stock policy. Prior ONNX release is preserved at tag pre-unity-safetensors-bundles-20260801.
