nx_clockjobs_gate.nx

nx_clockjobs_gate.nx -- is the scheduler's registry WELL-FORMED, or is it carrying silent no-ops? WHY THIS EXISTS (debt 1785601721). clk_load in nx_clock_sched.nx parses clock_jobs.tsv with NO validation of any kind. Every failure mode below is silent -- the row becomes a "job" that dispatches nothing, forever, and no instrument reports it: * fewer than 4 tab fields -> organ = "" -> dispatch is a no-op, scheduled for eternity * non-digit chars in a number -> SKIPPED, so an interval of "3o0" parses as 30 (wrong cadence, plausible value, no error) -- the silent-coercion class * interval = 0 -> a job that is always due * rows past CLK_MAXJOBS -> SILENTLY DROPPED (the truncating-cap class: the registry says one thing, the scheduler runs another) A MALFORMED ROW MUST BE REFUSED AT ADMISSION, NOT PARSED INTO A SILENT NO-OP AT DISPATCH. This gate is the DETECTION half of that: it cannot stop a bad write, but it makes one impossible to miss. The admission half (a validating writer) is the next rung and is deliberately separate -- a reader-side guard that is honest about being read-only beats a writer swap rushed under a live scheduler. SCOPE, STATED HONESTLY: this validates the REGISTRY FILE. It does NOT prove the named organ exists on disk, and it does NOT prove a job ever dispatched -- fail=0 is not evidence that a beat moved what it names. Those are separate claims and separate gates. license_tier: ORIGINAL expect_exit: 0