Most multimeters require the positive probe to be in a different socket in order to use the current range. I'd expect a 10A range and associated socket, and a few-hundred-milliamp range and another associated socket, along with the usual socket for measuring voltage, resistance, etc. The negative probe uses the same socket at all times. You'll probably need the 10A range to measure the short-circuit current of the dynamo.
Also, if the meter cost more than a fiver, there will be fuses on the current inputs (accessed via the battery compartment, of by taking the back off the meter). A standard multimeter rookie mistake is to attempt to measure the voltage of something capable of producing a decent amount of current (eg. a battery) with the probe still connected to the current socket. On a decent meter, this will immediately blow the fuse and leave you scratching your head at a reading of zero, before the realisation that you're going to have to fork out for some expensive new HRC fuses slowly dawns[1]. With a cheapo, unfused meter, this will briefly result in a nonsensical reading, while the realisation that the meter and/or the probe leads are in the process of catching fire slowly dawns[2].
To measure the resistance of the windings, you'll put the meter into the lowest value resistance mode (200Ω would be typical, if it doesn't auto-range), with the positive probe in the volts/ohms/etc socket, and probe the terminals of the stationary wheel. I'd expect a reading of about 3-6 ohms. The meter may helpfully beep while you do this (handy for working out what wire goes where, when you aren't fussed about the precise resistance). Coincidentally, this is the same procedure you'd use for determining whether a fuse is blown...
[1] You'll then forget all about it, and lend the meter to an electronics n00b trying to troubleshoot their dynamo, who will inexplicably get a reading of zero on the current range.
[2] I made this mistake on a car battery once. I realised as the smell of the melting probe lead reached my nose. Impressively, the two-for-a-fiver Maplin multimeter was undamaged.