Your pipeline can define its own environment variables and add them to the pipeline state. These will be available to all runs of that pipeline.
Use the utility functions for pipeline state management to add new environment variables to the current run state. For example, the following line in the execution block of a step:
add_pipeline_variables animal="dog" vegetable="carrot"
creates two environment variables available to all run of the pipeline where they were defined:
$ printenv animal dog $ printenv vegetable carrot
For more information on pipeline state, see Creating Stateful Pipelines.