diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 7b882ae5..457c4806 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -257,7 +257,7 @@ jobs:
             export S3_RAN=ran.synapse.org
           fi
 
-          curl -s https://raw.githubusercontent.com/Sage-Bionetworks/CI-Build-Tools/master/r-pkg/deploy.sh | bash
+          bash tools/deploy.sh
           # echo "::set-output name=s3_ran::$S3_RAN"
           echo "s3_ran=$S3_RAN" >> $GITHUB_OUTPUT
 
diff --git a/tools/deploy.sh b/tools/deploy.sh
new file mode 100644
index 00000000..5f818dbe
--- /dev/null
+++ b/tools/deploy.sh
@@ -0,0 +1,37 @@
+# This script takes as input the artifacts generated by build jobs and publishes the R packages
+# to a S3 hosted CRAN-like repository.  Users can then install the packages using the commmand:
+# install.packages(<package-name>, repos="http://<S3_RAN>.synapse.org")
+
+# Params
+# AWS_ACCESS_KEY_ID -- S3 creds
+# AWS_SECRET_ACCESS_KEY -- S3 creds
+# S3_RAN -- either "staging-ran.synapse.org" or "ran.synapse.org"
+# ARTIFACTS_DIR -- the directory where the artifacts are copied to
+
+home=`pwd`
+
+# remove the last build clone
+set +e
+rm -R ${S3_RAN}
+set -e
+
+# retrieving the current content of S3_RAN
+mkdir ${S3_RAN}
+cd ${S3_RAN}
+aws s3 sync s3://${S3_RAN}/ .
+cd ..
+
+curl -o deploy.R https://raw.githubusercontent.com/Sage-Bionetworks/CI-Build-Tools/master/r-pkg/deploy.R
+
+R -e "source('$home/deploy.R');\
+jenkins_deploy('$home/$S3_RAN', '$ARTIFACTS_DIR')"
+
+# upload
+cd ${S3_RAN}
+aws s3 sync --acl public-read . s3://${S3_RAN}/
+cd ..
+
+# This section is removed because gh-action
+# cleans up the home dir
+# clean up
+# rm -rf $home