diff --git a/jupyter/datascience/ubi8-python-3.8/setup-elyra.sh b/jupyter/datascience/ubi8-python-3.8/setup-elyra.sh
index 7443b6c69..1d96284f0 100644
--- a/jupyter/datascience/ubi8-python-3.8/setup-elyra.sh
+++ b/jupyter/datascience/ubi8-python-3.8/setup-elyra.sh
@@ -1,6 +1,39 @@
 #!/bin/bash
 set -x
 
+replace_invalid_characters (){
+  python -c 'import sys;print(sys.argv[1].translate ({ord(c): "-" for c in "!@#$%^&*()[]{};:,/<>?\|`~=_+"}))' "$1"
+}
+
+# Assumptions are existing kubeflow installation is in the kubeflow namespace
+DEFAULT_RUNTIME_FILE=$(jupyter --data-dir)/metadata/runtimes/test.json
+
+if [ -f "/var/run/secrets/kubernetes.io/serviceaccount/namespace" ]; then
+  SA_NAMESPACE=$(cat /var/run/secrets/kubernetes.io/serviceaccount/namespace)
+fi
+
+COS_BUCKET=$(replace_invalid_characters "$COS_BUCKET")
+export COS_BUCKET=${COS_BUCKET:-default}
+
+# If Kubeflow credentials are not supplied, use default Kubeflow installation credentials
+KF_DEPLOYMENT_NAMESPACE="${SA_NAMESPACE:=default}"
+AWS_ACCESS_KEY_ID="${AWS_ACCESS_KEY_ID:=minio}"
+AWS_SECRET_ACCESS_KEY="${AWS_SECRET_ACCESS_KEY:=minio123}"
+
+if [[ ! -f "$DEFAULT_RUNTIME_FILE" ]]; then
+  elyra-metadata install runtimes --schema_name=kfp \
+                                  --name=test \
+                                  --display_name=Test \
+                                  --auth_type=NO_AUTHENTICATION \
+                                  --api_endpoint=http://ml-pipeline."$KF_DEPLOYMENT_NAMESPACE".svc.cluster.local:3000/pipeline \
+                                  --cos_endpoint=http://minio-service."$KF_DEPLOYMENT_NAMESPACE".svc.cluster.local:9000 \
+                                  --cos_auth_type=USER_CREDENTIALS \
+                                  --cos_username="$AWS_ACCESS_KEY_ID" \
+                                  --cos_password="$AWS_SECRET_ACCESS_KEY" \
+                                  --cos_bucket="$COS_BUCKET" \
+                                  --engine=Tekton
+fi
+
 # Set the elyra config on the right path
 cp /opt/app-root/bin/utils/jupyter_elyra_config.py /opt/app-root/src/.jupyter/
 
@@ -12,3 +45,4 @@ fi
 # Environment vars set for accessing ssl_sa_certs and sa_token
 export KF_PIPELINES_SSL_SA_CERTS="/var/run/secrets/kubernetes.io/serviceaccount/ca.crt"
 export KF_PIPELINES_SA_TOKEN_ENV="/var/run/secrets/kubernetes.io/serviceaccount/token"
+export KF_PIPELINES_SA_TOKEN_PATH="/var/run/secrets/kubernetes.io/serviceaccount/token"
\ No newline at end of file
diff --git a/jupyter/datascience/ubi9-python-3.9/setup-elyra.sh b/jupyter/datascience/ubi9-python-3.9/setup-elyra.sh
index 7443b6c69..4da7bded1 100644
--- a/jupyter/datascience/ubi9-python-3.9/setup-elyra.sh
+++ b/jupyter/datascience/ubi9-python-3.9/setup-elyra.sh
@@ -1,6 +1,39 @@
 #!/bin/bash
 set -x
 
+replace_invalid_characters (){
+  python -c 'import sys;print(sys.argv[1].translate ({ord(c): "-" for c in "!@#$%^&*()[]{};:,/<>?\|`~=_+"}))' "$1"
+}
+
+# Assumptions are existing kubeflow installation is in the kubeflow namespace
+DEFAULT_RUNTIME_FILE=$(jupyter --data-dir)/metadata/runtimes/test.json
+
+if [ -f "/var/run/secrets/kubernetes.io/serviceaccount/namespace" ]; then
+  SA_NAMESPACE=$(cat /var/run/secrets/kubernetes.io/serviceaccount/namespace)
+fi
+
+COS_BUCKET=$(replace_invalid_characters "$COS_BUCKET")
+export COS_BUCKET=${COS_BUCKET:-default}
+
+# If Kubeflow credentials are not supplied, use default Kubeflow installation credentials
+KF_DEPLOYMENT_NAMESPACE="${SA_NAMESPACE:=default}"
+AWS_ACCESS_KEY_ID="${AWS_ACCESS_KEY_ID:=minio}"
+AWS_SECRET_ACCESS_KEY="${AWS_SECRET_ACCESS_KEY:=minio123}"
+
+if [[ ! -f "$DEFAULT_RUNTIME_FILE" ]]; then
+  elyra-metadata install runtimes --schema_name=kfp \
+                                  --name=test \
+                                  --display_name=Test \
+                                  --auth_type=NO_AUTHENTICATION \
+                                  --api_endpoint=http://ml-pipeline."$KF_DEPLOYMENT_NAMESPACE".svc.cluster.local:3000/pipeline \
+                                  --cos_endpoint=http://minio-service."$KF_DEPLOYMENT_NAMESPACE".svc.cluster.local:9000 \
+                                  --cos_auth_type=USER_CREDENTIALS \
+                                  --cos_username="$AWS_ACCESS_KEY_ID" \
+                                  --cos_password="$AWS_SECRET_ACCESS_KEY" \
+                                  --cos_bucket="$COS_BUCKET" \
+                                  --engine=Tekton
+fi
+
 # Set the elyra config on the right path
 cp /opt/app-root/bin/utils/jupyter_elyra_config.py /opt/app-root/src/.jupyter/
 
@@ -12,3 +45,4 @@ fi
 # Environment vars set for accessing ssl_sa_certs and sa_token
 export KF_PIPELINES_SSL_SA_CERTS="/var/run/secrets/kubernetes.io/serviceaccount/ca.crt"
 export KF_PIPELINES_SA_TOKEN_ENV="/var/run/secrets/kubernetes.io/serviceaccount/token"
+export KF_PIPELINES_SA_TOKEN_PATH="/var/run/secrets/kubernetes.io/serviceaccount/token"
diff --git a/jupyter/trustyai/ubi9-python-3.9/Pipfile b/jupyter/trustyai/ubi9-python-3.9/Pipfile
index c6d553561..a4d7b7b91 100644
--- a/jupyter/trustyai/ubi9-python-3.9/Pipfile
+++ b/jupyter/trustyai/ubi9-python-3.9/Pipfile
@@ -7,7 +7,7 @@ name = "pypi"
 
 [packages]
 # TrustyAI and useful extensions
-trustyai = "==0.2.12"
+trustyai = "~=0.2.12"
 
 # Datascience and useful extensions
 elyra-pipeline-editor-extension = "~=3.15.0"
diff --git a/jupyter/trustyai/ubi9-python-3.9/Pipfile.lock b/jupyter/trustyai/ubi9-python-3.9/Pipfile.lock
index 1b1d88e69..d9f37f7ac 100644
--- a/jupyter/trustyai/ubi9-python-3.9/Pipfile.lock
+++ b/jupyter/trustyai/ubi9-python-3.9/Pipfile.lock
@@ -1,7 +1,7 @@
 {
     "_meta": {
         "hash": {
-            "sha256": "160157c8ca0856da6dd43e10d773980e17d25497e426463458e9cf51385d5dc0"
+            "sha256": "336843943476abb02e536d5004c77d40c77fd6734da93b113ce8d73fb2e23c97"
         },
         "pipfile-spec": 6,
         "requires": {