Smoke test every finished deploy with Ghost Inspector

**Objective:** After the latest run of the GitHub workflow `{workflow-name}` (on branch `{branch}`, or any branch if blank), run the Ghost Inspector suite `{suite-id}` and report whether the deployed change passes.

**Step 1: Check the build (GitHub)**

Call `github-get-workflow-run` with `name` set to `{workflow-name}`, plus `branch` set to `{branch}` if given, and take the most recent run by `created-date`.

* If no run is found, report that and stop.
* If its `status` is not completed, report that it is still running and stop. Do not smoke test an unfinished deploy.
* If its `conclusion` is not success, report the run's conclusion and url and stop without running the suite.

**Step 2: Confirm the suite (Ghost Inspector)**

Call `ghost-inspector-get-suite` with `id` set to `{suite-id}`. If it is not found, stop. Tell the user which workflow run (run number, sha, branch) and which suite (name, test count, start url) you are about to run, and wait for confirmation.

**Step 3: Run it (Ghost Inspector)**

Call `ghost-inspector-execute-suite` with `suite-id` set to `{suite-id}`. Do not set `immediate`, so the call returns the result.

If the suite fails, call `ghost-inspector-get-test-result` with `suite-id` set to `{suite-id}`, `passing` false and `date-created.from` set to when you started the run, to name the failing tests.

**Step 4: Report**

* The workflow run: name, run number, branch, short sha, commit message and url.
* The suite: name, pass or fail, execution time, and the failing tests by name if any.
* One line verdict: the change at that sha passes or fails the smoke suite.

**Rules & Guidelines:**

* `github-get-workflow-run` is on the github server; the ghost-inspector tools are on the ghost-inspector server.
* `ghost-inspector-execute-suite` is the only write, called once, after confirmation.
* Do NOT call `github-rerun-workflow-run`, `github-cancel-workflow-run`, `github-add-issue` or `ghost-inspector-cancel-suite-run`. If the user wants an issue filed for a failure, say so and let them ask.

**Validation:**

* If the suite's start url does not look like the environment the workflow deploys to, warn the user before running and let them decide.
* If the execute call returns without a pass state, report the result id and tell the user the run is still in progress.