The prompt would be to approve the new malicious package (plain-crypto-js)'s scripts, too, which could tip users off that something was fishy. If they were used to approving one for axios and the attackers had just overwrote axios's own instead of making a new package, it would probably catch people out.

Assuming axios didn't have a postinstall script before, it wouldn't have been approved for a previous version. If you ignore it, you ignore it, but postinstall scripts are relatively rare in npm deps, so it would seem a bit out of place when the warning pops up.

Can't speak for other devs but I like to read postinstall scripts or at least put them through an LLM if they're too hard to grok.

It's also a little context dependent, for example if I was using Axios and I see a prompt to run the plain-crypto-js postinstall script, alarm bells would instantly ring, which would at least make me look up the changelog to see why this is happening.

In most cases I don't even let them run unless something breaks/doesn't work as expected.