Get a release go or hold from tests and open bugs

**Objective:** Decide whether the website is ready to release by running the Ghost Inspector suite `{suite-name}` (against `{start-url}` when given) and weighing the result with the open bugs in the BugHerd project `{project-name}`.

**Step 1: Find both**

Call `ghost-inspector-get-suite` with `name.contains` set to `{suite-name}`. If nothing matches, say so and stop. If several match, list their names and ids and ask which one. Call `bugherd-get-projects` with `name.contains` set to `{project-name}`. If nothing matches, say so and stop. If several match, list their names and project-id values and ask which one; do not guess.

**Step 2: Run the suite**

Tell the user which suite will run and against which URL, and wait for an explicit yes. Then call `ghost-inspector-execute-suite` once with `suite-id` (and `start-url` when given). Read the new run back with `ghost-inspector-get-suite-result` with `suite-id` and `max-records` 1, then the failing tests with `ghost-inspector-get-test-result` with `suite-id`, `passing` false and `date-created.from` set to the run's start.

**Step 3: Open bugs**

Call `bugherd-get-tasks` with `project-id`, once with `severity` critical and once with `severity` important. Drop done and closed tasks.

**Step 4: Verdict**

* **Hold:** any failing test, or any open critical bug.
* **Go with care:** all tests pass but important bugs are open; list them.
* **Go:** all tests pass and nothing critical or important is open.

**Step 5: Report**

The verdict first, then the run (passing, count-passing, count-failing, execution time), the failing tests with their first failed step (`ghost-inspector-get-test-step` with `test-result-id` and `passing` false), and the open bugs table.

**Rules & Guidelines:**

* Tools named bugherd-... are on the bugherd MCP server; tools named ghost-inspector-... are on the ghost-inspector MCP server. If either set of tools is missing, say which server is not connected and stop.
* The only write tool this prompt may call is `ghost-inspector-execute-suite`, once, after the user confirms. Do not create BugHerd tasks here; use `log-failing-tests-as-bugs` for that.

**Validation:**

* If the run is cancelled or errors, the verdict is "inconclusive"; say so.