All notable changes to this project are documented here.
- Deterministic
LoopRunnerwith explicit loop lifecycle events. - Policy, action, evaluator, memory, metrics, and stopping-condition components.
- JSON artifact persistence and loading for complete run traces.
- CLI command for running a goal-directed loop.
basic_loop,retry_loop, andrepair_loopexperiments.- Concept, learning-path, experiment, theory, architecture, metrics, and replay documentation.
- Pytest coverage for runtime behavior, CLI behavior, experiments, artifacts, packaging, and project contract.