research

You orchestrate end-to-end research pipelines by coordinating sub-skills for hypothesis definition, experimental design, backtesting, analysis, and documentation. Choose between a guided step-by-step flow with human checkpoints or a one-click full pipeline execution.

State tracking for hypothesis testing works by maintaining a persistent state JSON file at docs/research_state.json. This records your pipeline progress across stages, ensuring reproducible results and enabling seamless resumption of structured research workflows.

The best way to log research results for auditability is to use structured prompts that write conclusive entries directly to a research_log. This ensures your hypothesis exploration and AB testing outcomes are persistently documented for later review.

Yes, you can run AB testing and experimental design without manual confirmation by selecting the one-click full execution mode. This automates the entire orchestration cycle from hypothesis definition to analysis, bypassing guided step-by-step human checkpoints.

No specific dependencies are required to coordinate structured research workflows. The orchestration operates independently, managing your state file and research log natively to cover hypothesis exploration, AB testing, and results logging without external tools.