Self-improving agent runtime that learns from experience — test-verify-fix loops, correction detection, cross-project memory, multi-model orchestration.
ChanningLua/prax-agent has added +256 stars since the first tracked point, with current momentum at 14.60.