Specifically, I am planning to design and build an energy meter to measure power drawn from mains - 120V 60Hz. What are the risks associated with developing and testing this device and how to minimize/eliminate them?
This is such an important and useful topic that I'll post a quick subset answer to related issues now and try to get back for a much more complete answer later if it looks useful.
To start - a note on GFCIs and Isolating Transformers:
Isolating transformers and GFCIs both have potential (pun noted) value but also introduce new considerations.
GFCI = Ground Fault Current Interupter.
Also known as an ELCB and GFI and ... .
Current that is carried on both leads must be equal and opposite within some set limit.
If current supplied via the GFCI on phase does not return through the GFCI on neutral then the device disconnects power. Current trip limits are usually in the few 10's of mA range with a trip period shorter than the period required to cause ventricular fibrillation.
A "shock" via a GFCI as it trips will still be felt as a substantial electrical shock (ask me how I (purposefully) know), but it is brief.
A GFCI is very good for mains to true ground shock BUT if the "ground" contact is to the neutral side of the GFCI circuit then the user becomes "just part of the legitimate load" and there is zero protection.
ie When working on mains powered equipment a mains to neutral fault rather than mains to ground will not be protected against.
An isolating transformer is a 1:1 transformer with electrically isolated primary (mains input) winding and secondary (mains voltage) winding.
An isolating transformer has pros and cons.
With it you cannot sustain a mains to true-ground shock as there is no electrical path between any point on the secondary winding to mains-ground. (There may be minor unimportant coupling via inter-winding capacitance.) BUT
You can still sustain an isolated mains level voltage shock from isolated-phase to isolated system ground AND
Such a shock now wholly not protected by a mains side GFCI
See also this SE Q&A from 2011
How is using a transformer for isolation safer than directly connecting to the power grid?
Electrical shock than can obviously be via the heart "should be avoided" but any electrical shock has the potential to cause electrocution. eg a foot to foot shock MAY still cause electrocution.
Many substantial electric shocks do not kill (I've had a few over a long lifetime) but any one might. Arrange work practices and mental attitude to not have them.
Be wary of and look for situations that may lead to accidental shock - exposed conductors, casual attitudes, faulty equipment. (I had not had a major shock for maybe several decades - then a few years ago got caught by a faulty switchboard fuse holder that had missing insulation covering a phase-contacting screw - and conductive material had accumulated in the hole. I received a substantial hand to hand shock across the chest with substantial muscle soreness at the arm chest boundary on both arms due to muscular contraction.
DC shocks are potentially more dangerous than AC shocks as they can prevent muscular "let go" - under DC muscles clamp and do not release and it can be physically impossible to release one's grip.
With high conductivity skin contact a voltage of 12 Volts can be enough to cause almost unreleasable muscular contraction. (I have a friend who experienced this!. This is unusual - but all the more worth noting so as not to be 'caught out' in exceptional circumstances - Most people don't get to stand in seawater holding a metal handled flounder spear and a faulty LED 12V lantern mounted on a metal pole, as my friend did :-) ).
ELV DC systems at "safe" highly lethal voltages are increasingly common.
People have died from 36V helicopter battery contact.
Solar 48V systems are potential killers - but often an across the hand 48V shock is usually a mere nuisance 'tickle'. (During under-graduate experience training I worked on a 48V telecom wiring frame. In humid weather brushing the terminals of live circuits often resulted in annoying across hand mini shocks, with no other affects. Technicians in 48V powered exchanges worked on live equipment all the time with no warnings re shock hazard. 48V electrocution is rare. Don't be the exception.
In very worst case conditions 12 VDC shocks can kill. This can occur if voltage is applied across the chest with high conductive contact - unusual but not impossible (and has happened :-( - this occurred during an experiment when a prisoner agreed to participate in a medical experiment. Death was not intended, but happened. Microamp level currents under the skin & into the heart area can kill. [I'll add a reference 'later']. ).
With an AC shock it is often easier to release grip due to the cyclical zero crossings of current allowing muscles to repeatedly briefly cease contraction, but sometimes it is still "almost impossible" to release grip.