validating-real-scenarioslisted

# Validating real scenarios Tests cover the contracts someone knew to write down. A whole story driven through the shipped interfaces is what exposes the ones nobody knew about — the two components that each behave correctly and disagree about a format, the step that works in isolation and not third in a sequence. This skill owns which signal counts as a user-visible one. It does not judge what a green check proves in general, and it does not run a unit's exit gate. **"The suite is green" and "the product works" are different claims.** Only one of them has been observed at the end of a unit of work, and it is not the second one. ## Through the shipped interfaces, on a production-shaped instance `scenarios.stand_up`, `scenarios.suite` and `scenarios.permitted_entries` in `.ledger.yml` name how this project stands an instance up and what it drives. **Standing one up** stops on a missing value — say which one, because a scenario run against the wrong thing produces a green nobody can interpret. **Reading a suite that already exists stops on nothing.** Whether its steps go through a real entry, and whether its assertions could fail, is legible in the suite itself, and that reading is most of this skill. - **The instance is shaped like production, and is not production.** Same interfaces, same wiring, same startup path. - **Every step goes through a permitted entry.** A direct write to the store, a call into an internal function, a fixture that seeds state the product c