Replace k3s with direct soul-demo watchdog in Cloud Run container

Cloud Run gen2 doesn't provide eth0 with a unicast IP, causing k3s flannel
to crash on every container start. k3s was also wrong architecture for
Cloud Run (HPA inside a container, k3s overhead for one process).

Changes:
- entrypoint.sh: replace k3s server with a bash watchdog loop that starts
  soul-demo directly and restarts it on crash (3s backoff)
- Dockerfile.stage: remove k3s binary, soul-demo-image.tar, k3s manifests
  and their associated dirs/envvars; keep soul-demo binary only
- stage.yaml: remove 'Download k3s binary' step; rename and simplify
  soul-demo build step to compile binary only (no OCI image/tar)
- dev.yaml: update soul-demo placeholder step (binary not tar)
- manifest.el: document HAVE_CURL requirement since manifest.el has no
  c_flags/link_flags directive support

manifest.el

@@ -16,4 +16,12 @@ build {
     c_source "dist/page_css.c"
     c_source "dist/page_ga.c"
     c_source "dist/page_schema.c"
+    // NOTE: neuron-web requires el_runtime.c to be compiled with -DHAVE_CURL
+    // so that http_get/http_post forward to libcurl instead of returning
+    // {"error":"not built with HAVE_CURL"}. The elb binary in ci-base:dev
+    // hardcodes -DHAVE_CURL in its cc invocation, but older elb versions may
+    // not. manifest.el does not support c_flags or link_flags directives —
+    // if upgrading elb breaks HTTP, ensure ci-base:dev ships an elb built
+    // with HAVE_CURL enabled in its hardcoded cc command, or pre-compile
+    // el_runtime.o with -DHAVE_CURL on the host and pass it as a c_source.
 }