feat(ci): add CRD schema generation for kubeconform validation

Add scripts to extract OpenAPI v3 schemas from CRD definitions in all
kustomize overlays and write JSON schema files to ci/crd-schemas/ for
kubeconform validation. This allows kubeconform to validate CRD instances
(Elasticsearch, Kibana, CNPG Cluster, VictoriaMetrics, etc.) instead of
skipping or erroring on them.

- ci/generate-crd-schemas.py: extracts schemas from CRD YAML on stdin
- ci/generate-crd-schemas.sh: iterates overlays, pipes to Python script
- ci/validate-apps.sh, ci/validate-clusters.sh: add local schema-location fallback
- Makefile: add generate-schemas target
- add generate-schemas step to kubeconform woodpecker pipeline so schemas

ci/generate-crd-schemas.sh

@@ -0,0 +1,23 @@
+#!/usr/bin/env bash
+# Extract OpenAPI v3 schemas from CRD definitions in all kustomize overlays
+# and write JSON schema files to ci/crd-schemas/ for kubeconform validation.
+#
+# Run this script whenever CRD versions change, then commit the output.
+# Usage: ci/generate-crd-schemas.sh [output-dir]
+
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+OUTPUT_DIR="${1:-${SCRIPT_DIR}/crd-schemas}"
+mkdir -p "$OUTPUT_DIR"
+
+total=0
+
+while IFS= read -r -d "" k; do
+  dir="$(dirname "$k")"
+  n=$(kustomize build --enable-helm "$dir" 2>/dev/null \
+    | python3 "$SCRIPT_DIR/generate-crd-schemas.py" "$OUTPUT_DIR") || continue
+  total=$((total + n))
+done < <(find apps/overlays clusters -name kustomization.yaml -print0 | sort -z)
+
+echo "Generated ${total} schema(s) in ${OUTPUT_DIR}" >&2