A significant proportion of cheap USB chargers are defective by design with minimal isolation between the mains supply and their low voltage output. e.g:
If you measure the output ripple and noise, without using an isolation transformer for its mains supply, sooner or later you will encounter one that's got defective insulation, shorted between its primary and secondary sides which will blow up your scope. Its possible to use a CAT II or better rated high voltage differential probe or isolated differential probe, to avoid the need for an isolation transformer, but such probes are expensive, and have significantly lower bandwidth than even an ordinary cheap x10 probe, so unless you intend to take measurements on the primary side of mains PSUs or other circuits that have no mains isolation, (for which differential probes are vastly preferable
*), I would recommend an isolation transformer. Note that the isolation transformer must be wired for test bench use with no ground connection on the secondary side.
It would also be desirable to do a 1KV
HiPot test (or PAT test) on all the chargers to eliminate those that have already failed (or are about to fail) dangerously.
With that out the way, you'll need to be able to tap USB Vbus and Gnd right at the charger output. Others have suggested a few options for doing so.
Also note that cheap SMPSUs can interact badly with active loads (e.g. the control loops of a SMPSU and OPAMP+ MOSFET current sink can be unstable when connected together resulting in oscillation), so even if you use an active load for most measurements, its highly desirable to have suitable power resistors for a purely resistive load, (and preferably a power rheostat to vary the load resistance) so you can be certain the noise/ripple is from the charger and not its interaction with an active load.
* Novice and less experienced technicians should *NOT* attempt to make oscilloscope measurements on any sort of non-isolated mains circuit (e.g. the primary side of a SMPSU) without appropriate training and a formal risk assessment, even with a differential probe.