refactor(server): drop HTTP /cron route; run crons in-process only

The in-process scheduler is the sole cron trigger on self-host, so the
/cron/{name} HTTP endpoint and its CRON_SHARED_SECRET are dead surface.
Reduce the HTTP server to GET / (health) and remove the cron secret config,
SSM binding, and now-unused timeout.

internal/modules/cron_dispatcher.go

@@ -6,16 +6,16 @@ import (
 	"fmt"
 )
 
-// ErrCronNotFound is returned when /cron/{name} addresses an unregistered cron.
+// ErrCronNotFound is returned when name addresses an unregistered cron.
 var ErrCronNotFound = errors.New("cron not found")
 
 // DispatchScheduled runs the cron registered under name with the per-module
 // prefixed Deps the registry stored at Build time. Returns ErrCronNotFound if
-// no module owns that name — EventBridge Scheduler hitting an unknown route is a
-// configuration bug worth surfacing as a 404 at the HTTP layer.
+// no module owns that name — a scheduler firing an unknown name is a
+// configuration bug worth surfacing to the caller.
 //
 // The handler runs synchronously in the calling goroutine; ctx propagates
-// cancellation/timeout from the HTTP request.
+// the scheduler's per-fire cancellation/timeout.
 func DispatchScheduled(ctx context.Context, name string, reg *Registry) error {
 	cron, ok := reg.Cron(name)
 	if !ok {