Surface colour alone is an unreliable indicator of airframe material and this is especially true of stealth aircraft, where there are many more layers of coatings than typical. I absolutely expect the J-20 to have a very substantial composite content despite the uniformly yellow paint.
OTOH this works both ways, the belief that everything painted in green primer on Chinese aircraft is automatically made out of composite is equally misguided. Consider this picture of a Rafale undergoing final assembly:
You could be forgiven for thinking that it supports the view that green means composite, as major composite parts (wing skins, rear fuselage skins, fin skins, forward fuselage skins) are indeed green while virtually all parts known to be metal are yellow. All the same, several parts which are definitely composite are also yellow (rudder, ailerons, elevons).
Then there are these Airbus A320-family airframes:
All of the fuselage is known to be metal, yet various parts are green while others are yellow, because they are supplied by plants all over Europe which may use different primer formulations (so long as it meets the spec, it matters not). Not to mention the yellow fin torque box and rudder on the first one, both of which are actually CFRP parts. And while we're on the subject of Airbus:
The vast majority of the A350 fuselage is composite (and still it is yellow), the only part which is confirmed to be metal is... the cockpit area, which is green.
In the absence of good sources like the Rafale material break-down graphic you have to exercise some judgement on whether the material distribution suggested by surface colour makes sense. On SAC's Flankers for example it frequently does not - I doubt their composite content is nearly as high as is frequently suggested.
EDIT: Yeah, what taxiya says, basically. The Y-20 is indeed another good example.