Require the smallest real test before expanding an experiment

Before expanding an experiment harness or reusable architecture, require the smallest real-world test that can distinguish the hypothesis. Define the decision, baseline, minimum sample, primary metric, guardrails, budget, and stop/promote rule first. If the real test cannot run, pause the expansion and record the evidence and exact resumption condition. Code volume, test volume, sessions, or token use show allocation, not success. This guardrail prevents strong implementation work from outgrowing the unresolved question. It does not mean broad infrastructure is wasteful; platform work can proceed when reuse or risk reduction is itself the predeclared outcome.

Applies when: exploratory product, model, or architecture work can expand before live validation. Keywords: experiment gate smallest test stop rule budget validation outcome