A summary of all sandbox verdicts for a given file. It's a dictionary, where each key is the sandbox name and each value is a dictionary containing the following keys:

category: <string> normalized verdict category. It can be one of suspicious, malicious, harmless or undetected.
confidence: <integer> verdict confidence from 0 to 100.
malware_classification: <list of strings> raw sandbox verdicts.
malware_names: <list of strings> malware family names.
sandbox_name: <string> sandbox that provided the verdict.