fix(builtins): prevent awk parser panic on multi-byte UTF-8

The awk parser used byte offsets with chars().nth() (char index),
causing panics when multi-byte UTF-8 appeared in comments, strings,
or regex patterns. Added current_char()/advance() helpers for safe
char-boundary handling and replaced all chars().nth(byte_offset)
patterns with byte-safe slicing.

Closes #395

Author: claude