AA2 character data format

I'll try to detail some of my findings of the AA2 character data file format here. I did most of the work charting this info myself, but some fellow anons in /hgg/ helped out initially so big thanks to them.
The data is stored inside a PNG container, and all offsets are based of the first IEND (49 45 4E 44) of the file. All values are in hex (0x) unless otherwise specified.
| Offset | What                 | Length | Value/Info
+--------+----------------------+--------+-----------------------------------------
|      0 | PNG IEND             | 4      | (49 45 4E 44)
|     1D | Last name text       | 104    | Null terminated, so effectively 259 characters max. Text format info below.
|    121 | First name text      | 104    | Same as above.
|    225 | Profile text         | FF     | Editor says max 255 chars, but keeps reading if not terminated.
|    692 | Intelligence         | 1      | 00 to 04: Lowest -> Highest.
|    698 | Strength             | 1      | 00 to 04: Lowest -> Highest.
|    6A3 | Sociability          | 1      | 00 to 04: Lowest -> Highest.
|    6A4 | Fighting style       | 1      | 00 to 02: No restraint, One blow, Irregular.
|    6A5 | Virtue               | 1      | 00 to 04: Lowest -> Highest.
|    6A6 | Sexual orientation   | 1      | 00 to 04: Hetero, Lean hetero, Bisexual, Lean homo, Homo.
|    6A7 | Sexual experience    | 1      | 00 or 01 (False or True)
|    6A8 | Anal experience      | 1      | 00 or 01
|    6AA | Character traits     | 26     | There are 38 traits each one byte, value is either 00 or 01. Order same as in the editor.
|    6D0 | Rainbow status       | 1      | 00 or 01
|    6D1 | Sexual Preferences   | D      | Same as the traits, but only 13.
¤ These are the confirmed values for now. I assume most of the other follows roughly in the same order they are presented in the character editor, as that seems to be the common pattern. I will edit in more values once I have confirmed them properly.
Text format
------------
The text format was a real pain to figure out. The text is coded in japanese Shift-JIS encoding, and then all the bytes are flipped for some weird reason I really don't understand. So to read it back you first need to extract a text block, then do a bitwise inversion on all bytes (xor FF or ~). You then get Shift-JIS (where ASCII is readable at least) which can be converted to something actually useful (UTF-8). To write back you just reverse the process. Also note that since this seems to be using C-strings you should terminate all text with null (\0) before inverting (or FF (~\0) after the invert).