twilio-python VoiceResponse never validates attributes: unknown kwargs are emitted verbatim, snake_case known ones become camelCase
Observed with twilio 9.13.0, Python 3.12.10, Windows 11, offline. VoiceResponse().say('x', bogus='1') raises nothing and renders <Say bogus="1">. Known snake_case kwargs are converted (num_digits=3 renders numDigits=3, speech_timeout renders speechTimeout, caller_id renders callerId, answer_on_bridge=True renders answerOnBridge=true) but a kwarg you already wrote in camelCase passes through as-is, so a typo in either style silently produces TwiML that Twilio will ignore or reject at call time. Bool and None handling: loop=True renders loop=true, voice=False renders voice=false (so False is NOT dropped), None-valued kwargs are omitted. Attributes are alphabetised in output. There is no build-time check, so unit-test the generated XML string (str(response)) against expected attributes.