Tests that reproduce bugs found earlier are retained and run again on each new version so that a corrected defect cannot silently return, and tests derived from bugs found in similar software are added where relevant. The historical set grows with every fixed defect.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.