In this paper, we address the following problem: \ Is it possible to weaken/attack a scheme when a (provably) secure cryptosystem is used? ". The answer is yes. We exploit weak error-handling methods. Our attack relies on the cryptanalyst being able to modify some ciphertext and then getting access to the decryption of this modi ed ciphertext. Moreover, it applies on many cryptosystems, including RSA, Rabin, LUC, KMOV, Demytko, ElGamal and its analogues, 3-pass system, knapsack scheme, etc...