Answer two questions that get confused constantly: is this image soft, and did my own pipeline soften it. Symptom is judged only by same-size 1:1 eyes-on comparison (three automated sharpness metrics were measured, all three wrong); responsibility is judged by per-pixel diff against the generator's original output, never by a metric. Includes the 'original files are listed in your own deliverable script' rule, a don't-try-again list of three failed metrics with their exact wrong numbers, and two zero-dependency scripts (same-head-width comparator, pixel-diff forensics).

Install

openclaw skills install @catsmansion/mv-image-forensics