Limitcode 1.1.7 - Literal Unicode escapes preserved (issue #9)

edit_file and write_to_file no longer rewrite literal \uXXXX sequences.
Content like {"arrow": "\u2192"} in JSON, JavaScript or regexes was
silently converted to the actual character, changing the written file.

Escapes are now decoded only when a tool match proves the provider
corrupted the arguments (seen with some providers), so broken-provider
repair keeps working while legitimate escape text is left untouched.
