US Patent 10445501 Detecting malicious scripts

Systems and methods normalize an executable script. A file can be received that potentially contains an executable script. The characters in the file are translated to a single case (either upper case or lower case). Duplicate whitespace can be removed. A script is identified within the file. Tokens in the script are processed to create normalized output. The normalized output can include tokens that are retained keywords, control flow characters or data characters from the script file.