switched model back to gpt-4o-mini due to poor performance / adding ytb videos due to block

changed transcription library

increased max steps to 15

submission version

added additional libraries

added comment

added builtins in authorized imports

improved .py tool, final_answer and changed model to gpt-4.1-mini

removed command to delete file audio.mp3

add MockResponse class to simulate response.content

Initial commit with LFS-tracked files

initial commit