refactor(runner): caller manages app launch/terminate

Removes Launch + Terminate from runner.Run so the CLI can launch
the app first, wait for the SDK to connect, then start the loop.
The previous shape forced runner to launch internally which fought
with the SDK-must-be-connected-first ordering.

BundleID/ClearState fields go away too since runner no longer
launches; the CLI keeps them on its testOptions struct.

internal/runner/runner.go

@@ -15,8 +15,6 @@ import (
 )
 
 type Options struct {
-	BundleID        string
-	ClearState      bool
 	Duration        time.Duration
 	SnapshotTimeout time.Duration
 	IdleTimeout     time.Duration
@@ -39,20 +37,16 @@ type ViolationRecord struct {
 	Properties []string
 }
 
+// Run drives the snapshot/evaluate/release/act loop until the duration
+// elapses or the context is canceled. The caller is responsible for
+// launching the app and connecting the SDK before Run is called, and for
+// terminating the app afterwards.
 func Run(ctx context.Context, options Options) (Summary, error) {
 	if err := validate(options); err != nil {
 		return Summary{}, err
 	}
 
 	summary := Summary{StartTime: time.Now()}
-
-	if err := options.Driver.Launch(ctx, options.BundleID, options.ClearState); err != nil {
-		return summary, fmt.Errorf("launch: %w", err)
-	}
-	defer func() {
-		_ = options.Driver.Terminate(context.Background())
-	}()
-
 	deadline := summary.StartTime.Add(options.Duration)
 	stepIndex := 0
 	for time.Now().Before(deadline) {