Comment by arcfour

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.