The cube string is a 54 character string that encodes the state of a 3x3 cube.

There are 9 characters for each face, following the order URFDLB, and within a face following the diagram below.

             +------------+
             | U1  U2  U3 |
             |            |
             | U4  U5  U6 |
             |            |
             | U7  U8  U9 |
+------------+------------+------------+------------+
| L1  L2  L3 | F1  F2  F3 | R1  R2  R3 | B1  B2  B3 |
|            |            |            |            |
| L4  L5  L6 | F4  F5  F6 | R4  R5  R6 | B4  B5  B6 |
|            |            |            |            |
| L7  L8  L9 | F7  F8  F9 | R7  R8  R9 | B7  B8  B9 |
+------------+------------+------------+------------+
             | D1  D2  D3 |
             |            |
             | D4  D5  D6 |
             |            |
             | D7  D8  D9 |
             +------------+

A mask is a 54 character string of only 'x' and '.', where 'x' is a sticker that will be greyed out and '.' is a sticker that will be kept unchanged.

Initial masks are applied to the initial solved cube before it is scrambled, like removing the stickers on a physical cube.
The grey stickers can be moved.
They can be used to train a substep. e.g. removing all of the side top layer stickers to train OLL, or removing all top layer stickers except the top cross to train ZBLS.

Final masks are applied to the cube that is already scrambled into a training case.
The grey stickers cannot be moved.
They can be used to block vision of parts of the cube in order to practice limited side recognition such as 2-sided PLL recognition.

Program to generate the mask string from a list of stickers