providing, to an artificial intelligence (AI) model, a malicious script that implements a malicious behavior, wherein the AI model is configured to modify software code of the malicious script to produce modified software code that obfuscates the malicious behavior;

producing, by a processing device using the AI model and based on the malicious script, an adversarial script comprising the modified software code that obfuscates the malicious behavior;

determining whether the adversarial script and the malicious script are functionally equivalent in performing the malicious behavior; and

in response to determining that the adversarial script and the malicious script are functionally equivalent in performing the malicious behavior, adding the adversarial script to a validated adversarial scripts database that is utilized to test a malware detector.