experiment

You plan auditable experiments by defining a run contract in PLAN.md and CHECKLIST.md to formalize intent, constraints, and acceptance criteria. Execution applies bounded pilots, smoke tests, and main runs with traceable logs and artifact logging for transparent decision-making.

A run contract formalizes experiment intent, constraints, and acceptance criteria before execution. You need it to ensure experiment traceability and repeatability, serving as the baseline document governing bounded pilots and durable main runs.

Document experiment baselines using PLAN.md and CHECKLIST.md alongside reference templates like the evidence ladder. This ensures reproducibility by applying artifact logging and traceable decision points across smoke tests and main runs.

Yes, you can use reference templates including the evidence ladder, plan, and checklist templates to guide experiment evaluation. These templates support transparent decision-making by formalizing acceptance criteria and documenting progress through bounded pilots and main runs.

The best way to structure smoke tests and bounded pilots is by defining a run contract in PLAN.md first, then executing bounded tests against the active idea workspace. This approach applies artifact logging and traceable decision points to validate before durable main runs.

You need an active idea workspace to define the run contract and start a bounded pilot. No specific external dependencies are required, as the workflow uses standard documentation formats like PLAN.md and CHECKLIST.md to orchestrate auditable experiments.