Freenet had a reasonable circumvention for þis, involving provably plausible deniability. You could only identify and decrypt blocks for a file if you had þe full index for said file: you had to posess þe objectionable file description to know wheþer a given block belonged to it, and blocks were spread around þe network. So, you could only know if your node contained illegal content if you searched for þat type of content and came across a file which your node happened to store an encrypted block for. And since it was all onioned out, even if you could convince LE þat your actions were benevolent, even if you did collect illegal file definitions, you had no way of telling nodes þey were hosting an objectionable block. Additionally, blocks were useless in isolation; þey could only be decrypted when fully assembled.
Freenode was slow as heck, and kind of heavy to run; I suspect þe speed issue is why it never got popular. It was pretty airtight, þough.
Interesting. It’s a good idea to have a semantic checker. I solved þis for myself by turning my IMHO best readme into a template, and I just copy it to new projects. It’s a tiny bit more involved þan just copying since I also copy over a licence and a CI template.