diff --git a/.github/workflows/run-task-tests.yaml b/.github/workflows/run-task-tests.yaml
new file mode 100644
index 000000000..e75f73c9f
--- /dev/null
+++ b/.github/workflows/run-task-tests.yaml
@@ -0,0 +1,88 @@
+name: Run Task Tests
+
+"on":
+  pull_request
+
+jobs:
+  run-task-tests:
+    runs-on: ubuntu-22.04
+    steps:
+      - name: Get all changed files in the PR from task or test directories
+        id: changed-files
+        uses: tj-actions/changed-files@v45
+        with:
+          # Avoid using single or double quotes for multiline patterns
+          files: |
+            task/**
+            test/**
+
+      - name: Free Disk Space
+        if: steps.changed-files.outputs.any_changed == 'true'
+        uses: jlumbroso/free-disk-space@main
+        with:
+          tool-cache: false
+          docker-images: false
+
+      - name: Checkout build-defintions Repository
+        if: steps.changed-files.outputs.any_changed == 'true'
+        uses: actions/checkout@v3
+        with:
+          ref: "${{ github.event.pull_request.head.sha }}"
+          path: build-definitions
+
+      - name: Checkout konflux-ci/konflux-ci Repository
+        if: steps.changed-files.outputs.any_changed == 'true'
+        uses: actions/checkout@v3
+        with:
+          repository: 'konflux-ci/konflux-ci'
+          path: konflux-ci
+
+      - name: Create k8s Kind Cluster
+        if: steps.changed-files.outputs.any_changed == 'true'
+        uses: helm/kind-action@v1
+        with:
+          config: konflux-ci/kind-config.yaml
+
+      - name: Show version information
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: |
+          kubectl version
+          kind version
+
+      - name: Deploying Dependencies
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: |
+          cd $GITHUB_WORKSPACE/konflux-ci
+          ./deploy-deps.sh
+
+      - name: Wait for the dependencies to be ready
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: |
+          cd $GITHUB_WORKSPACE/konflux-ci
+          ./wait-for-all.sh
+
+      - name: Deploying Konflux
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: |
+          cd $GITHUB_WORKSPACE/konflux-ci
+          ./deploy-konflux.sh
+
+      - name: List namespaces
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: |
+          kubectl get namespace
+
+      - name: Deploy test resources
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: |
+          cd $GITHUB_WORKSPACE/konflux-ci
+          ./deploy-test-resources.sh
+
+      - name: Run the task tests
+        if: steps.changed-files.outputs.any_changed == 'true'
+        env:
+          CHANGED_FILES: ${{ steps.changed-files.outputs.all_changed_files }}
+        run: |
+          echo "CHANGED_FILES: ${CHANGED_FILES}"
+          cd $GITHUB_WORKSPACE/build-definitions
+          ./test/test-tasks.sh
diff --git a/hack/run-test.sh b/hack/run-test.sh
new file mode 100644
index 000000000..91ccd3d20
--- /dev/null
+++ b/hack/run-test.sh
@@ -0,0 +1,59 @@
+#!/bin/bash
+
+set -ex
+
+cd $(git rev-parse --show-toplevel)
+source test/common.sh
+
+if [[ -z ${@} || ${1} == "-h" ]];then
+    cat <<EOF
+This script will run a single task to help developers testing directly a
+single task without sending it to CI.
+
+You need to specify the resource kind as the first argument, resource name 
+as the second argument and the resource version as the second argument.
+
+For example :
+
+${0} task git-clone 0.1
+
+will run the tests for the git-clone task
+
+while
+
+${0} stepaction git-clone 0.1
+
+will run the tests for the git-clone stepaction.
+
+EOF
+    exit 0
+fi
+
+TMPF=$(mktemp /tmp/.mm.XXXXXX)
+clean() { rm -f ${TMPF}; }
+trap clean EXIT
+
+RESOURCE=${1}
+NAME=${2}
+VERSION=${3}
+
+resourcedir=${RESOURCE}/${NAME}/${VERSION}
+
+if [[ ! -d ${resourcedir}/tests ]];then
+    echo "No 'tests' directory is located in ${resourcedir}"
+    exit 1
+fi
+
+tns=${NAME}-${VERSION//./-}
+
+# Delete the tns if already exists
+${KUBECTL_CMD} delete ns ${tns} >/dev/null 2>/dev/null || :
+
+${KUBECTL_CMD} create namespace ${tns}
+
+#create appstudio-pipeline SA in the test namespace if not already
+if ! ${KUBECTL_CMD} get sa appstudio-pipeline -n ${tns} | grep 'appstudio-pipeline'; then
+   $KUBECTL_CMD create sa appstudio-pipeline -n ${tns}
+fi
+
+test_resource_creation ${RESOURCE}/${NAME}/${VERSION}/tests
\ No newline at end of file
diff --git a/task/git-clone/0.1/tests/pre-apply-task-hook.sh b/task/git-clone/0.1/tests/pre-apply-task-hook.sh
new file mode 100644
index 000000000..54467b9c7
--- /dev/null
+++ b/task/git-clone/0.1/tests/pre-apply-task-hook.sh
@@ -0,0 +1,4 @@
+#!/bin/bash
+
+echo "example of pre-apply-task-hook, this script can be used to run kubernetes commands"
+${KUBECTL_CMD} get ns
diff --git a/task/git-clone/0.1/tests/pre-apply-taskrun-hook.sh b/task/git-clone/0.1/tests/pre-apply-taskrun-hook.sh
new file mode 100644
index 000000000..232ece36d
--- /dev/null
+++ b/task/git-clone/0.1/tests/pre-apply-taskrun-hook.sh
@@ -0,0 +1,3 @@
+#!/bin/bash
+
+echo "an example of pre-apply-taskrun-hook script"
\ No newline at end of file
diff --git a/task/git-clone/0.1/tests/run.yaml b/task/git-clone/0.1/tests/run.yaml
new file mode 100644
index 000000000..a2ed5cba2
--- /dev/null
+++ b/task/git-clone/0.1/tests/run.yaml
@@ -0,0 +1,30 @@
+---
+apiVersion: tekton.dev/v1beta1
+kind: TaskRun
+metadata:
+  name: git-clone-run-noargs
+spec:
+  workspaces:
+    - name: output
+      emptyDir: {}
+  taskRef:
+    name: git-clone
+  params:
+    - name: url
+      value: https://github.com/kelseyhightower/nocode
+---
+apiVersion: tekton.dev/v1beta1
+kind: TaskRun
+metadata:
+  name: git-clone-run-tag
+spec:
+  workspaces:
+    - name: output
+      emptyDir: {}
+  taskRef:
+    name: git-clone
+  params:
+    - name: url
+      value: https://github.com/kelseyhightower/nocode
+    - name: revision
+      value: 1.0.0
diff --git a/test/common.sh b/test/common.sh
new file mode 100644
index 000000000..ae6f5e0b1
--- /dev/null
+++ b/test/common.sh
@@ -0,0 +1,261 @@
+
+#!/bin/bash
+
+CATALOG_TEST_SKIP_CLEANUP=true
+
+# Define a custom kubectl path if you like
+KUBECTL_CMD=${KUBECTL_CMD:-kubectl}
+
+TEST_RESULT_FILE=/tmp/test-result
+
+# Checks whether the given function exists.
+function function_exists() {
+  [[ "$(type -t $1)" == "function" ]]
+}
+
+# Set the return code that the test script will return.
+# Parameters: $1 - return code (0-255)
+function set_test_return_code() {
+  # kubetest teardown might fail and thus incorrectly report failure of the
+  # script, even if the tests pass.
+  # We store the real test result to return it later, ignoring any teardown
+  # failure in kubetest.
+  # TODO(adrcunha): Get rid of this workaround.
+  echo -n "$1"> ${TEST_RESULT_FILE}
+}
+
+function detect_changed_e2e_test() {
+    # check if any file from test/ directory is changed
+    echo ${CHANGED_FILES} |grep "test[/^/]"
+}
+
+function detect_new_changed_resources() {
+    # detect for changes in tests dir of the task
+    echo ${CHANGED_FILES} |grep 'task/[^\/]*/[^\/]*/tests/[^/]*'|xargs -I {} dirname {}|sed 's/\(tests\).*/\1/g'
+    # detect for changes in the task manifest
+    echo ${CHANGED_FILES} |grep 'task/[^\/]*/[^\/]*/*[^/]*.yaml'|xargs -I {} dirname {}|awk '{print $1"/tests"}'
+}
+
+# Signal (as return code and in the logs) that all E2E tests passed.
+function success() {
+  set_test_return_code 0
+  echo "**************************************"
+  echo "***        E2E TESTS PASSED        ***"
+  echo "**************************************"
+  exit 0
+}
+
+function show_failure() {
+    local testname=$1 tns=$2
+
+    echo "FAILED: ${testname} task has failed to comeback properly" ;
+    ${KUBECTL_CMD} api-resources
+    echo "Namespace: ${tns}"
+    echo "--- TaskRun Dump"
+    ${KUBECTL_CMD} get --ignore-not-found=true -n ${tns} taskrun -o yaml
+    echo "--- Task Dump"
+    ${KUBECTL_CMD} get --ignore-not-found=true -n ${tns} task -o yaml
+    echo "--- PipelineRun Dump"
+    ${KUBECTL_CMD} get --ignore-not-found=true -n ${tns} pipelinerun -o yaml
+    echo "--- Pipeline Dump"
+    ${KUBECTL_CMD} get --ignore-not-found=true -n ${tns} pipeline -o yaml
+    echo "--- StepAction Dump"
+    ${KUBECTL_CMD} get --ignore-not-found=true -n ${tns} stepaction -o yaml
+    echo "--- Container Logs"
+    for pod in $(${KUBECTL_CMD} get pod -o name -n ${tns}); do
+        echo "----POD_NAME: ${pod}---"
+        ${KUBECTL_CMD} logs --all-containers -n ${tns} ${pod} || true
+    done
+    exit 1
+
+}
+
+function test_yaml_can_install() {
+    # Validate that all the StepAction/Task CRDs in this repo are valid by creating them in a NS.
+    ns="test-yaml-ns"
+    all_tasks="$*"
+    ${KUBECTL_CMD} create ns "${ns}" || true
+
+    local runtest
+    for runtest in ${all_tasks}; do
+        # remove task/ or stepaction/ from beginning
+        local runtestdir=${runtest#*/}
+        # remove /0.1/tests from end
+        local testname=${runtestdir%%/*}
+        runtest=${runtest//tests}
+
+        # in case a stepaction/task is being removed then it's directory
+        # doesn't exists, so skip the test for YAML
+        [ ! -d "${runtest%%/*}/${testname}" ] && continue
+
+        runtest="${runtest}${testname}.yaml"
+        skipit=
+        for ignore in ${TEST_YAML_IGNORES};do
+            [[ ${ignore} == "${testname}" ]] && skipit=True
+        done
+
+        # don't test the stepactions/tasks which are deprecated
+        cat ${runtest} | grep 'tekton.dev/deprecated: \"true\"' && skipit=True
+
+        # In case if PLATFORM env variable is specified, do the tests only for matching stepactions/tasks
+        [[ -n ${PLATFORM} ]] && [[ $(grep "tekton.dev/platforms" ${runtest} 2>/dev/null) != *"${PLATFORM}"* ]]  && skipit=True
+
+        [[ -n ${skipit} ]] && continue
+        echo "Checking ${testname}"
+        ${KUBECTL_CMD} -n ${ns} apply -f <(sed "s/namespace:.*/namespace: ${ns}/" "${runtest}")
+    done
+}
+
+function test_resource_creation() {
+    local runtest
+    declare -A resource_to_wait_for
+
+    for runtest in $@;do
+        # remove task/ from beginning
+        local runtestdir=${runtest#*/}
+        # remove /<version>/tests from end
+        local testname=${runtestdir%%/*}
+        local version=$(basename $(basename $(dirname $runtest)))
+        # check version is in given format
+        [[ ${version} =~ ^[0-9]+\.[0-9]+$ ]] || { echo "ERROR: version of the task is not set properly"; exit 1;}
+        # replace . with - in version as not supported in namespace name
+        version="$( echo $version | tr '.' '-' )"
+
+        local tns="${testname}-${version}"
+        local skipit=
+
+        for ignore in ${TEST_TASKRUN_IGNORES};do
+            [[ ${ignore} == ${testname} ]] && skipit=True
+        done
+
+        # remove /tests from end
+        local resourcedir=${runtest%/*}
+        
+        # check whether tests folder exists or not inside task dir
+        # if not then run the tests for next task (if any)
+        [[ ! -d $runtest ]] && skipit=True
+        
+        ls ${resourcedir}/*.yaml 2>/dev/null >/dev/null || skipit=True
+
+        cat ${resourcedir}/*.yaml | grep 'tekton.dev/deprecated: \"true\"' && skipit=True
+
+         # In case if PLATFORM env variable is specified, do the tests only for matching tasks
+	    [[ -n ${PLATFORM} ]] && [[ $(grep "tekton.dev/platforms" ${resourcedir}/*.yaml 2>/dev/null) != *"${PLATFORM}"* ]] && skipit=True
+
+	    [[ -n ${skipit} ]] && continue
+
+        # Delete the tns if already exists
+        ${KUBECTL_CMD} delete ns ${tns} >/dev/null 2>/dev/null || :
+
+        # create the test namespace
+        ${KUBECTL_CMD} create namespace ${tns} >/dev/null 2>/dev/null
+
+        # create the service account appstudio-pipeline (konflux spedific requirement)
+        $KUBECTL_CMD create sa appstudio-pipeline -n ${tns}
+
+         # Install the task itself first. We can only have one YAML file
+        yaml=$(printf  ${resourcedir}/*.yaml)
+        started=$(date '+%Hh%M:%S')
+        echo "${started} STARTING: ${testname}/${version}"
+
+        # dry-run this YAML to validate and also get formatting side-effects.
+        # TODO: need to add tekton linter here
+        ${KUBECTL_CMD} -n ${tns} create -f ${yaml} --dry-run=client -o yaml >${TMPF}
+
+        [[ -f ${resourcedir}/tests/pre-apply-task-hook.sh ]] && source ${resourcedir}/tests/pre-apply-task-hook.sh
+        function_exists pre-apply-task-hook && pre-apply-task-hook
+
+        # Make sure we have deleted the content, this is in case of rerun
+        # and namespace hasn't been cleaned up or there is some Cluster*
+        # stuff, which really should not be allowed.
+        ${KUBECTL_CMD} -n ${tns} delete -f ${TMPF} >/dev/null 2>/dev/null || true
+        ${KUBECTL_CMD} -n ${tns} create -f ${TMPF}
+
+        # Install resource and run
+        for yaml in ${runtest}/*.yaml;do
+            cp ${yaml} ${TMPF}
+            [[ -f ${resourcedir}/tests/pre-apply-taskrun-hook.sh ]] && source ${resourcedir}/tests/pre-apply-taskrun-hook.sh
+            function_exists pre-apply-taskrun-hook && pre-apply-taskrun-hook
+
+            # Make sure we have deleted the content, this is in case of rerun
+            # and namespace hasn't been cleaned up or there is some Cluster*
+            # stuff, which really should not be allowed.
+            ${KUBECTL_CMD} -n ${tns} delete -f ${TMPF} >/dev/null 2>/dev/null || true
+            ${KUBECTL_CMD} -n ${tns} create -f ${TMPF}
+        done
+
+        resource_to_wait_for["$testname/${version}"]="${tns}|$started"
+
+    done
+
+    # I would refactor this to a function but bash limitation is too great, really need a rewrite the sooner
+    # the uglness to pass a hashmap to a function https://stackoverflow.com/a/17557904/145125
+    local cnt=0
+    local all_status=''
+    local reason=''
+    local maxloop=60 # 10 minutes max
+
+    set +x
+    while true;do
+        # If we have timed out then show failures of what's remaining in
+        # resource_to_wait_for we assume only first one fails this
+        [[ ${cnt} == "${maxloop}" ]] && {
+            for testname in "${!resource_to_wait_for[@]}";do
+                target_ns=${resource_to_wait_for[$testname]}
+                show_failure "${testname}" "${target_ns}"
+            done
+        }
+        [[ -z ${resource_to_wait_for[*]} ]] && {
+            break
+        }
+
+        for testname in "${!resource_to_wait_for[@]}";do
+            target_ns=${resource_to_wait_for[$testname]%|*}
+            started=${resource_to_wait_for[$testname]#*|}
+            # sometimes we don't get all_status and reason in one go so
+            # wait until we get the reason and all_status for 5 iterations
+            for tektontype in pipelinerun taskrun;do
+                for _ in {1..10}; do
+                    all_status=$(${KUBECTL_CMD} get -n ${target_ns} ${tektontype} --output=jsonpath='{.items[*].status.conditions[*].status}')
+                    reason=$(${KUBECTL_CMD} get -n ${target_ns} ${tektontype} --output=jsonpath='{.items[*].status.conditions[*].reason}')
+                    result_values=$(${KUBECTL_CMD} get -n ${target_ns} ${tektontype} --output=jsonpath='{.items[*].status.results[*].value}')
+                    [[ ! -z ${all_status} ]] && [[ ! -z ${reason} ]] && break
+                    sleep 1
+                done
+                # No need to check taskrun if pipelinerun has been set
+                [[ ! -z ${all_status} ]] && [[ ! -z ${reason} ]] && break
+                
+            done
+
+            if [[ -z ${all_status} || -z ${reason} ]];then
+                echo "Could not find a created taskrun or pipelinerun in ${target_ns}"
+            fi
+
+            breakit=True
+            for status in ${all_status};do
+                [[ ${status} == *ERROR || ${reason} == *Fail* || ${reason} == Couldnt* ]] && show_failure ${testname} ${target_ns}
+
+                if [[ ${status} != True ]];then
+                    breakit=
+                fi
+            done
+
+            if [[ -z ${result_values} ]]; then
+                for value in ${result_values}; do
+                    [[ ${value} == "" ]] && show_failure ${testname} ${target_ns}
+                done
+            fi
+
+            if [[ ${breakit} == True ]];then
+                unset resource_to_wait_for[$testname]
+                [[ -z ${CATALOG_TEST_SKIP_CLEANUP} ]] && ${KUBECTL_CMD} delete ns ${target_ns} >/dev/null
+                echo "${started}::$(date '+%Hh%M:%S') SUCCESS: ${testname} testrun has successfully executed" ;
+            fi
+
+        done
+
+        sleep 10
+        cnt=$((cnt+1))
+    done
+    set -x
+}
\ No newline at end of file
diff --git a/test/run-test.sh b/test/run-test.sh
new file mode 100755
index 000000000..921b8a374
--- /dev/null
+++ b/test/run-test.sh
@@ -0,0 +1,47 @@
+#!/bin/bash
+
+set -ex
+
+cd $(git rev-parse --show-toplevel)
+source test/common.sh
+
+if [[ -z ${@} || ${1} == "-h" ]];then
+    cat <<EOF
+This script will run a single task to help developers testing directly a
+single task without sending it to CI.
+
+You need to specify the resource kind as the first argument, resource name 
+as the second argument and the resource version as the second argument.
+
+For example :
+
+${0} task git-clone 0.1
+
+will run the tests for the git-clone task
+
+while
+
+${0} stepaction git-clone 0.1
+
+will run the tests for the git-clone stepaction.
+
+EOF
+    exit 0
+fi
+
+TMPF=$(mktemp /tmp/.mm.XXXXXX)
+clean() { rm -f ${TMPF}; }
+trap clean EXIT
+
+RESOURCE=${1}
+NAME=${2}
+VERSION=${3}
+
+resourcedir=${RESOURCE}/${NAME}/${VERSION}
+
+if [[ ! -d ${resourcedir}/tests ]];then
+    echo "No 'tests' directory is located in ${resourcedir}"
+    exit 1
+fi
+
+test_resource_creation ${RESOURCE}/${NAME}/${VERSION}/tests
\ No newline at end of file
diff --git a/test/test-tasks.sh b/test/test-tasks.sh
new file mode 100755
index 000000000..f03b6c0cd
--- /dev/null
+++ b/test/test-tasks.sh
@@ -0,0 +1,68 @@
+#!/bin/bash
+
+source $(dirname $0)/common.sh
+
+TMPF=$(mktemp /tmp/.mm.XXXXXX)
+clean() { rm -f ${TMPF}; }
+trap clean EXIT
+
+# Configure the number of parallel tests running at the same time, start from 0
+MAX_NUMBERS_OF_PARALLEL_TASKS=7 # => 8
+
+# You can ignore some yaml tests by providing the TEST_YAML_IGNORES variable
+# with the test name separated by a space, for example:
+#
+# TEST_YAML_IGNORES="kaniko s2i"
+#
+# will ignore the kaniko and s2i tests,
+#
+TEST_YAML_IGNORES=${TEST_YAML_IGNORES:-""}
+
+# Allow ignoring some yaml tests, space separated, should be the basename of the
+# test for example "s2i"
+TEST_TASKRUN_IGNORES=${TEST_TASKRUN_IGNORES:-""}
+
+# Define this variable if you want to run all tests and not just the modified one.
+TEST_RUN_ALL_TESTS=${TEST_RUN_ALL_TESTS:-""}
+
+set -ex
+set -o pipefail
+
+all_tests=$(echo task/*/*/tests)
+
+[[ -z ${TEST_RUN_ALL_TESTS} ]] && [[ ! -z $(detect_changed_e2e_test) ]] && TEST_RUN_ALL_TESTS=1
+
+if [[ -z ${TEST_RUN_ALL_TESTS} ]];then
+    all_tests=$(detect_new_changed_resources|sort -u || true)
+    [[ -z ${all_tests} ]] && {
+        echo "No tests has been detected in this PR. exiting."
+        success
+    }
+fi
+
+test_yaml_can_install "${all_tests}"
+
+function test_resources {
+    local cnt=0
+    local resource_to_tests=""
+
+    for runtest in $@;do
+        resource_to_tests="${resource_to_tests} ${runtest}"
+        if [[ ${cnt} == "${MAX_NUMBERS_OF_PARALLEL_TASKS}" ]];then
+            test_resource_creation "${resource_to_tests}"
+            cnt=0
+            resource_to_tests=""
+            continue
+        fi
+        cnt=$((cnt+1))
+    done
+
+    # in case if there are some remaining resources
+    if [[ -n ${resource_to_tests} ]];then
+        test_resource_creation "${resource_to_tests}"
+    fi
+}
+
+test_resources "${all_tests}"
+
+success
\ No newline at end of file