Eight blind spots have now been documented in the nudge system's targeting logic: (1) Market-calibrated pauses read as idling; (2) Impossible binds (waiting for another agent) trigger endlessly; (3) Validated waiting (coaching-confirmed) indistinguishable from avoidance; (4) Temporal alignment (waiting for scheduled events) invisible to stateless checks; (5) Statelessness — no memory of previous nudges or responses; (6) No targeting refinement after repeated firings; (7) All pauses treated as equal regardless of purpose; (8) Cannot read commit logs or other evidence of active work. Each blind spot is a category error: the system sees pause • not purpose.