It is a general consensus that physical access can be dangerous. However, I would like to know attack vectors if attackers only get nearby access to the system.
I’m assuming that the word access refers to contact. Sometimes, attackers may not be able to physically contact the machine, but may be able to get close to it.
For example, when I’m taking a bus and put my tablet into my bag, an attacker can get very close to it without acting suspicious, especially when the bus is full of passengers. When I’m using my computer in a (public) library, the guy sitting in front of the next desk may be within 1 meter from my computer.
I’m aware of several attack vectors, but I’m not sure whether there are others I haven’t come up with, and I may underestimate the risk. Those include:
Looking at my keyboard input. While I can use other means of authentication (e.g. biometrics), sometimes they are not available (I believe most computers don’t, at least without specialised hardware).
Motion sensors that record oscillations. The sensors may eventually gather enough information to crack passwords, etc.
Insecure transfer protocol. If there is a bug within the protocol e.g. AirDrop, nearby attackers may exploit that bug and trigger a buffer overflow. Unencrypted transfers like those under http may be read or intercepted by attackers.
I don’t think these are very useful attacks when the attackers cannot get physical access afterwards, and some of them can be easily mitigated (e.g. securely store passwords in browsers, enable 2FA, always use VPN). Not so sure whether there are other things that attackers can do, though. Any help is appreciated.