Claude Code agent for rigorous ML hypothesis investigation: critic-defender debate (default) or ensemble critique (opt-in), pre-specified metrics, agreed empirical tests, and production re-evaluation. Methodology itself is empirically calibrated through pre-registered meta-experiments.
machine-learning hypothesis-testing peer-review ml-agents llm agentic-workflow claude-code-plugin empirical-validation research-workflow ensemble-critique
-
Updated
May 25, 2026 - Python