Overview

Currently, when content is encrypted and shattered and packets are written, only specific directories are modified based on the random seed. However, this means an attacker can check the last modified timestamp of each directory in storage. For example, if only 2 directories were modified within a minute when the file was encrypted, the attacker can know the packets are within those two directories and narrow his attack and search to those two directories only. The problem becomes bigger when there are many more subdirectories. Even if you have a hundred subdirectories, if only 5-10 are modified, the system is easier to crack.

Possible fixes

1. Touch all directories. You could either do this by writing a junk file atleast n numbers in each subdirectory in such a way that they are virtually identical to obscure which directories contain the files.

Bonus bounty points for ensuring time obscurity while doing this (ensuring attacker cannot know which packets are junk and which are real by looking at timestamps of modifying/creation. )

2. Modify the system to distribute existing files across nearly/all directories in a reversible/retrievable manner. (not recommended)