update skills docs

.claude/skills/trace-manifestwork/SKILL.md

@@ -43,6 +43,15 @@ This skill traces these relationships bidirectionally, combining database querie
 
 ## Key Concepts
 
+### Terminology
+
+**IMPORTANT**: In Maestro colloquial usage, **ManifestWork** and **resource bundle** are the same concept and are used interchangeably:
+
+- **ManifestWork**: The formal Kubernetes Custom Resource Definition (CRD) name used by the Open Cluster Management SDK
+- **Resource bundle**: The term used in Maestro's RESTful API endpoints (e.g., `/api/maestro/v1/resource-bundles`)
+
+When users refer to "resource bundles," they are talking about ManifestWork resources. Both terms describe a collection of Kubernetes manifests packaged together for delivery to target clusters. The database stores these in the `resources` table, while the Kubernetes cluster manages them as ManifestWork/AppliedManifestWork CRs.
+
 ### Cluster Architecture
 
 **CRITICAL**: Maestro uses a dual-cluster architecture:

.claude/skills/trace-manifestwork/scripts/trace.sh

@@ -182,13 +182,32 @@ query_db() {
         echo ""
     fi
 
-    # Check for postgres-breakglass (ARO-HCP INT - CRITICAL ENVIRONMENT)
+    # Check for maestro-db first (normal database pod)
+    pod_name=$(kubectl_svc -n maestro get pods -l name=maestro-db -o jsonpath='{.items[0].metadata.name}' 2>/dev/null || echo "")
+
+    if [[ -n "$pod_name" ]]; then
+        echo "Database: maestro-db"
+        echo ""
+
+        # Try to detect the postgres user
+        local db_user
+        db_user=$(kubectl_svc -n maestro exec -i "$pod_name" -- env 2>/dev/null | grep '^POSTGRES_USER=' | cut -d= -f2)
+
+        if [[ -z "$db_user" ]]; then
+            db_user="maestro"
+        fi
+
+        kubectl_svc -n maestro exec -i "$pod_name" -- psql -U "$db_user" -d maestro -c "$sql_query"
+        return 0
+    fi
+
+    # Fallback: Check for postgres-breakglass (ARO-HCP INT - CRITICAL ENVIRONMENT)
     local breakglass_deployment
     breakglass_deployment=$(kubectl_svc -n maestro get deployment postgres-breakglass -o jsonpath='{.metadata.name}' 2>/dev/null || echo "")
 
     if [[ -n "$breakglass_deployment" ]]; then
+        echo "Database: postgres-breakglass (fallback - maestro-db not found)"
         echo "Environment: ARO-HCP INT (CRITICAL)"
-        echo "Database: postgres-breakglass"
         echo ""
 
         # Check if pod is running
@@ -257,16 +276,6 @@ query_db() {
         return 0
     fi
 
-    # Check for maestro-db (Service cluster)
-    pod_name=$(kubectl_svc -n maestro get pods -l name=maestro-db -o jsonpath='{.items[0].metadata.name}' 2>/dev/null || echo "")
-
-    if [[ -n "$pod_name" ]]; then
-        echo "Database: maestro-db (Service cluster)"
-        echo ""
-        kubectl_svc -n maestro exec -i "$pod_name" -- psql -U maestro -d maestro -c "$sql_query"
-        return 0
-    fi
-
     echo "ERROR: No database pod found on service cluster"
     echo "Checked for:"
     echo "  - postgres-breakglass deployment (ARO-HCP INT)"

.claude/skills/trace-resource-request/SKILL.md

@@ -35,6 +35,16 @@ Use this skill when:
 
 **Common scenario**: You know a deployment or service isn't working, but only know its Kubernetes name. Use `trace-manifestwork` to find the associated resource ID and work name, then use this skill to trace through server and agent logs to identify where the request failed or stalled.
 
+## Terminology
+
+**IMPORTANT**: In Maestro colloquial usage, **ManifestWork** and **resource bundle** are the same concept and are used interchangeably:
+
+- **ManifestWork**: The formal Kubernetes Custom Resource Definition (CRD) name used by the Open Cluster Management SDK
+- **Resource bundle**: The term used in Maestro's RESTful API endpoints (e.g., `/api/maestro/v1/resource-bundles`)
+- **Resource**: The term used in database tables (`resources` table) and some internal code
+
+When users refer to "resource bundles," "resources," or "ManifestWorks," they are all talking about the same thing: a collection of Kubernetes manifests packaged together for delivery to target clusters. In log messages and traces, you'll see all three terms used to refer to this concept.
+
 ## Request Flow Overview
 
 Understanding the complete request flow is critical for effective troubleshooting: