Dangerzone (PDF Sandboxing)

[Average_Joe (Joey)]

@beantaco

The 4.5mb PDF converting to a 70mb PDF is unsurprising. As you may know, the reason the created PDF is much larger than the original PDF is probably rasterization. An image of a US-letter or A4 page that has enough resolution to be readable is huge in file size, and there would be multiple pages of them in a single document. Unfortunately AFAIK there is no way to adjust the resolution, compression level, color channels etc.

I appreciate your reply!

Is using Dangerzone the best way to make sure that PDF files are safe to be opened?

Is there another option? Shouldn’t it be enough to just scan the PDF file with an anti-virus scanner?

[Expert4870]

@Average_Joe

Shouldn’t it be enough to just scan the PDF file with an anti-virus scanner?

What threats are you trying to mitigate? It seems it would take a determined adversary targeting dangerzone specifically.

If it’s a zero day than I assume it wouldn’t be detected by anti virus and they face false positives and it’s not a good approach because it’s badness enumeration.

I assume combining a secure PDF viewer like GrapheneOS pdf viewer and DangerZone would be a reasonably secure approach.

On desktop you could use a chromium brower (or whatever the most secure desktop pdf browser is) in a qube and dangerzone.

[szp2005]

Worth adding to the rasterizing point: it also throws out the text layer. After Dangerzone the page is just pixels, so anything selectable or searchable has to be rebuilt by OCR, and what OCR gives you is a reconstruction, not the original text. It misreads characters and scrambles tables. If you quote from documents or grep them, that cost bites harder than the file size.