How do I work out if a USB-C cable is safe to use?

There is no need to suspect something or not. Everything is defined in Type-C specifications. Use the common rule of engineering: Read specifications.

Yes, the main problem is with “legacy cables”, Type-A plug to Type-C in particular. The Type-C specifications define the primary mechanism of how the consuming port (phone/tablet, or Upstream Facing Port, UFP) detects source capability.

If a standard C-C cable is used, the supplying port (Downstream Facing port, DFP) “advertises” its capability by using three different pull-up resistors on CC pin. If a 5V pull-up reference is used, these values are 56k, 22k, and 10k, for port’s capability of 500mA, 1500mA, and 3000mA correspondingly. The CC wire propagates this information from DFP to the end of Type-C cable. The connected device (phone) will (should) detect this, and limit its consumption accordingly.

Now, what to do if you have only a legacy Type-A port on your host, as most PCs do? The Type-A does not have any extra pins like CC. The Type-C Specification suggests embedding this information into the Type-C end of the legacy A->C cable. Therefore, the “information channel” is broken now, and the phone will try to grab as much current as the pull-up resistor inside the Type-C overmold indicates, which is soldered by cable’s manufacturer. Since the cable does not know which port you will plug it in, the safe cable pull-up should be 56k, otherwise the phone can try to suck 1.5A or even 3A from the cable. If the port is a regular USB, the requested cable power may vastly exceed port's capability. With cheap uncontrolled power delivery (some cheap PC mainboards connect the VBUS directly to internal +5VSTBY), it will cause system shutdown.

If the port is powerful enough, but the cable in use is skinny (Type-C cables can have as low as 28AWG on VBUS and GND wires) and the C-connector has wrong 10k pull-up, the cable might burn out and cause fire.

P.S. You can measure the pull-up value of CC-to VBUS on any A-C cable by using a breakout connector like this one: link

A guide on how to buy safe USB-C cables was prepared by laptopmag.com and is available here:

http://www.laptopmag.com/articles/how-to-find-safe-usb-type-c-cables

Suffice to say that testing currents and resistances from connector to connector is going to be nontrivial and would likely require specialized gear. Trust, then, in cables certified by people who have that gear. As an example, looking at the engineer you mention in your post, Benson Leung's, google+ blog, you can get an idea about his testing methodology. A lot of it seems to be a combination of plug-and-pray and putting things under an oscilloscope.

edit: There is a diagram of how the resistors are arranged within a USB-C cable available here. It may be possible to use a USB-C breakout board, such as the one available here, to establish baseline impedances and resistances across terminals of known-good (certified) cables and test those baselines against suspicious cables. Be sure to account for tolerances.

I did a bit of homework on the subject - I don't have the equipment to definitively check for safety yet but there's a handful of things I noticed with my own experiments.

For most part, this issue seems to be with USB C to USB 2.0 type A cables. They misidentify themselves to certain phones - the nexuses and pixels seem to be the primary ones, and draw significantly more power than is safe. C to C cables should be fine.

That said, the phone, computer and cable all seem to play a part.

If you have a nexus 5x, 6p or chromebook pixel, the checkR app would be a good way to do a quick test. I don't and this doesn't work on my device. These seem to be the primary devices mentioned in most articles on the matter and are both the best tools to check and the devices you need to exercise the most caution with.

I use ampere instead since I have a One Plus three - it gives me a "max usb current reading" (which I only see on my USB C capable phone and not my old moto G)- and on a dedicated powerbank or charger, its 1500mA, and on front panel USB ports its detected as 500mA.

A "safe" cable will have a value under 3A and of course, current draw should be under that unless you're using a fast charger - the DASH/VOOC charger/cable combination is detected as 1800mA USB max, and charges at 3500mA (As designed). I believe ampere shows the phone's charge rate so your actual current draw would be approximately 200-500mA more. Current monitor would give you raw data as well.

I suspect the OP3 drops down to 1.5A charging (and 500mA on a PC)unless the DASH/VOOC cable/charger is used so it might be safe by default by virtue of being out of specification. If you're unsure, its probably worth keeping an eye on it.

I may buy a dedicated USB volt/ammeter for science in future but it seems that unless your phone's designed to draw more current based on the cable as per the spec it should be fairly safe.

So, My phone seems to behave well with any cable, and my cheap dodgy cables are safe for most part. If anyone visits with a nexus, I'll get them to run checkR first, and annotate my cables.