Justifiably so, based on past known threat models, the crypto world strongly advices in favor of hardware wallets, but a recent somehow unrelated thread got me thinking of a different concern than what the original intent of hardware wallets was: what is the likelihood of a combination between internal network access of hardware wallets manufacturers and their devices assembly source code repository (e.g. local version control systems), or private github repositories, or even third party publicly available code libraries, if used? Of course brute forcing keys is practically impossible, but insertion of malicious code during the image installs or updates of wallets feels to me a more likely scenario of concern.