Maybe the ransomware author used someone else's library instead of implementing textbook RSA from scratch.
It also occurs to me that (unfortunately) the ransomware setting may be one where comparatively few kinds of attacks are feasible. The ransomware will encrypt one short fixed-length random value (chosen by itself, not the user) once and then stop. The public key is presumably fixed and was most likely generated offline using a separate tool like OpenSSL.
The decryption presumably happens only on the ransomware author's infrastructure and is gated by a payment, so it's potentially hard to perform oracle attacks (and perhaps different kinds of decryption failures don't produce meaningfully different observable behavior, especially if a human being is in the loop returning the decryption tokens to the ransomware victims who've paid the random).
It also occurs to me that (unfortunately) the ransomware setting may be one where comparatively few kinds of attacks are feasible. The ransomware will encrypt one short fixed-length random value (chosen by itself, not the user) once and then stop. The public key is presumably fixed and was most likely generated offline using a separate tool like OpenSSL.
The decryption presumably happens only on the ransomware author's infrastructure and is gated by a payment, so it's potentially hard to perform oracle attacks (and perhaps different kinds of decryption failures don't produce meaningfully different observable behavior, especially if a human being is in the loop returning the decryption tokens to the ransomware victims who've paid the random).