Interesting approach. I've spent a lot of time wrangling PDF internals recently, and the issue is usually maintaining the xref table integrity when you inject new content streams.
Does this approach rewrite the entire file structure on save, or are you appending incremental updates to the EOF? Incremental is safer for corruption, but file size bloats quickly with AI-generated diffs.