# TEST-01 coverage experiment

This fixture tests one narrow claim from the TEST-01 evidence brief: two test
suites can report equal structural coverage while only the suite containing
the business boundary kills an off-by-one mutation at that boundary.

Run it from this directory:

```sh
docker build --tag serein-test-01 .
docker run --rm --volume "$PWD/results:/experiment/results" serein-test-01
```

The container replaces its files inside `results/`, then records tool
versions, coverage reports, mutation logs, commands, durations, and exit
codes. The committed `composer.lock` and image digests pin the dependency
graph and base images.

The experiment is illustrative. It is not evidence about a production system
or about all test suites. A result is publishable only when `comparison.txt`
confirms equal coverage and the logs show the selected mutation escaping Suite
A and being killed by Suite B.