The recovered files were malformed. Claude Code helped work out that each was a metadata header, a thumbnail and the real image stuck together, then wrote a script that split them, metadata intact.
Result Way faster than doing it by hand
“I had it write a python script and was able to recover all the images with their metadata.”
Spotted: found in the wild. Details are as reported by the source above and have not been independently verified.
Recovered files from a failing SD card with GetDataBack, but the images were malformed and wouldn’t open. Had Claude Code analyse the files, and they worked out that each one was a custom metadata header, a thumbnail and the actual image joined together. Then had it write a Python script to pull them apart.
They recovered all the images with their metadata. Nothing a human couldn’t have figured out, they say, but way faster than doing it themselves.
I recovered files from a damaged SD card, but they won't open: [describe them; attach a few samples]. Analyse the file structure byte by byte, compare it with the normal [JPEG or RAW] format, and explain what's wrong. Then write a Python script that extracts the real images, with their metadata if possible, from every file in [folder] into [output folder] without touching the originals.
A starting prompt written by Agents in the Wild. Fill in the [brackets], and check what the agent does before it sends, pays for or deletes anything.