Add /health endpoint for aggregated skill status

scripts/run.sh

@@ -68,27 +68,93 @@ serve_metrics() {
     exec socat TCP-LISTEN:$METRICS_PORT,reuseaddr,fork EXEC:"/tmp/metrics_handler.sh"
 }
 
+# Create health check script
+create_health_script() {
+    cat > /tmp/health.sh << 'SCRIPT'
+#!/bin/bash
+SKILLS_DIR="${SKILLS_DIR:-/skills}"
+
+# Health check endpoints for known skills
+declare -A HEALTH_PORTS=(
+    ["caddy"]="2019"
+    ["ttyd"]="7681"
+    ["loki"]="3100"
+    ["duckdb"]="8432"
+    ["claude"]="8888"
+    ["metrics"]="9090"
+)
+
+result='{"status":"healthy","skills":{'
+first=true
+
+for skill_dir in "$SKILLS_DIR"/*/; do
+    skill_name=$(basename "$skill_dir")
+    [ "$skill_name" = "supervisor" ] && continue
+    [ "$skill_name" = "metrics" ] && continue
+
+    port="${HEALTH_PORTS[$skill_name]}"
+    [ -z "$port" ] && continue
+
+    # Check health
+    if curl -sf --max-time 2 "http://localhost:$port/health" >/dev/null 2>&1 || \
+       curl -sf --max-time 2 "http://localhost:$port/" >/dev/null 2>&1; then
+        status="up"
+    else
+        status="down"
+    fi
+
+    $first || result+=','
+    first=false
+    result+="\"$skill_name\":\"$status\""
+done
+
+result+='}}'
+
+# Check if any are down
+if echo "$result" | grep -q '"down"'; then
+    result=$(echo "$result" | sed 's/"healthy"/"degraded"/')
+fi
+
+echo "$result"
+SCRIPT
+    chmod +x /tmp/health.sh
+}
+
 # Create HTTP handler script
 create_handler_script() {
     cat > /tmp/metrics_handler.sh << 'HANDLER'
 #!/bin/bash
-# Read HTTP request (we ignore it, always serve /metrics)
+# Read HTTP request
 read -r request_line
+path=$(echo "$request_line" | cut -d' ' -f2)
+
 while read -r header; do
     [ "$header" = $'\r' ] && break
 done
 
-# Get metrics
+send_response() {
-metrics=$(/tmp/aggregate.sh)
+    local content_type="$1"
-content_length=${#metrics}
+    local body="$2"
+    local body_length=${#body}
 
-# Send HTTP response
+    printf "HTTP/1.1 200 OK\r\n"
-printf "HTTP/1.1 200 OK\r\n"
+    printf "Content-Type: %s\r\n" "$content_type"
-printf "Content-Type: text/plain; charset=utf-8\r\n"
+    printf "Content-Length: %d\r\n" "$body_length"
-printf "Content-Length: %d\r\n" "$content_length"
+    printf "Connection: close\r\n"
-printf "Connection: close\r\n"
+    printf "\r\n"
-printf "\r\n"
+    printf "%s" "$body"
-printf "%s" "$metrics"
+}
+
+case "$path" in
+    /health)
+        health=$(/tmp/health.sh)
+        send_response "application/json" "$health"
+        ;;
+    /metrics|*)
+        metrics=$(/tmp/aggregate.sh)
+        send_response "text/plain; charset=utf-8" "$metrics"
+        ;;
+esac
 HANDLER
     chmod +x /tmp/metrics_handler.sh
 }
@@ -103,5 +169,6 @@ trap cleanup SIGTERM SIGINT
 
 start_node_exporter
 create_aggregator_script
+create_health_script
 create_handler_script
 serve_metrics