exp-design

To convert a research hypothesis into structured experiment blocks, you map claims into baseline, validation, ablation, and robustness phases. This creates dependency-ordered execution steps with explicit success criteria and budgets.

Claim-driven experiment design for robustness testing isolates and stresses potential threats to a hypothesis. It structures experiments to verify target claims, decompose variables, and refute invalid assumptions with reproducible steps.

You can generate ablation studies from a wiki idea slug by running an automated experiment design pass. It reads the slug, scopes the target claims, and outputs isolation blocks that test individual factor contributions.

Yes, you can catch missing baselines and weak statistical rigor by running a reviewer pass. This review analyzes the experimental plan to identify unfair comparisons and propose concrete fixes before execution.

To add tested_by graph edges linking claims to experiments, you write generated experiment pages under the wiki directory and rebuild graph context artifacts. This creates structural links connecting verified claims to their execution blocks.

The best way to organize baseline and validation experiments for reproducibility is structuring them into ordered blocks with explicit success criteria. Baselines reproduce existing results while validation blocks prove the target hypothesis.