run-scenariolisted

# Run a Scenario One protocol on both ends of the ODD loop: the same commands that produced the numbers in an observation report produce the numbers that verify the fix. A scenario that cannot be replayed verbatim makes before/after comparison an impression, not a measurement. **A step this package ships a script for is run, never rewritten.** The replay of a stored benchmark is `k6-guides`' script; the machine preflight and the service probe are their own skills'. Authoring a shell script that redoes one of them costs turns before it runs and yields a different command each time — two runs then measure two things. ## Read by situation **The two mechanical parts are scripts, not prose.** Replaying a stored benchmark is `scripts/replay_benchmark.py` (see [`references/benchmark-replay.md`](references/benchmark-replay.md)), and pointing the query CLI at the local stack is the `setup-local-stack` skill's — its `## Configure an isolated context` ships it as one command. Run them; do not rebuild them. What is left in these files is what you have to decide, which is the only part worth reading. This file is the method every scenario follows — steps 1 to 5 and the rules. What depends on the situation lives in a reference, read by the block that applies, never whole: | Situation | Reference | | --- | --- | | Every drive: the clean-base order and the identity the queries are qualified by, and the run's t0 after the warmup — a process the run launches, a port already served, a re