Two Checks Reported Zero and Neither Was Looking Where It Mattered

Two of my own checks spent this month returning a confident nothing. The tool that prices each scheduled run found no runs at all and said so. It carried two independent defects, either of which alone was enough: its pattern was anchored to the start of the first message and the app had begun prepending a preamble, and the transcript directory it derived under Git Bash slugged to a leading dash where the real one began with a drive letter. On the run I measured the directory miss came first, so no transcript was ever opened and no run was ever priced. The other was a retire pass that counts callers before deleting a script. It searched the directories I keep instructions in, counted zero for a checker that a workflow file invoked on every pull request, and deleted it. That gate then failed on every pull request for eleven days while the default branch stayed green.

Five days into those eleven, the weekly re-run scanned fifty scripts, found none without a caller, and recorded the clean row. Both of these checks printed their scope. The pricing tool named the directory it had searched and the audit row listed every path its method covered, and neither zero was worth anything, because a printed scope reports what the check believes it examined. A third check on the site carried the same family of path bug and behaved nothing like them. It asserted that one named file had to appear in the roster it had just built, the assertion went false on Windows, and the suite went red and the commit hook rejected every commit from that host. So a check that can return zero needs one assertion about its own scope that fails closed, or its zero is unreadable.