Currently when you upgrade the plugin you are at risk that new reporters are added/integrated and these collide with existing pipeline code you have to parse these results.

This leads to duplicate reports.

On top of this you may want to get the version of the thing you are building by useing something like

withMaven() {
  bsVersion = returnStdout: true, script: sh "mvn help:evaluate -Dexpression=project.version | grep -E '^[[:digit:]+.]+.*-SNAPSHOT' "
}

unfortunately using the above means that the taskScanner is currently invoked (which will then be invoked multiple times) which will cause unnecessary IO, fingerprinting is also involved - which again when you just want to run a simple maven command with no scanners is not good.  The workaround of adding 

withMaven(options: [concordionPublisher(disabled: true), findbugsPublisher(disabled: true), dependenciesFingerprintPublisher(disabled: true), artifactsPublisher(disabled: true), invokerPublisher(disabled: true), junitPublisher(disabled: true, ignoreAttachments: false), openTasksPublisher(disabled: true), pipelineGraphPublisher(disabled: true)]) {
// some block
}

is both verbose and will likely not do what is expected when a new publisher/reporter is integrated.