A method for scanning files for security, including receiving an unfamiliar file for scanning, if the determining indicates that the mime type is suitable for analysis, then processing a buffer of file data from the unfamiliar file, including generating a histogram of frequencies of occurrence of bytes within a buffer of file data from the unfamiliar file, excluding a designated set of bytes, and if the generated histogram of frequencies of occurrence of the non-excluded bytes deviates substantially from a reference distribution, then signaling that the unfamiliar file is potentially malicious. A system and a computer-readable storage medium are also described and claimed.