Artifical photosynthesis has been making a lot of progress lately. Similar to your concept of a layer of skin which converts chemical energy to electricity, a possible better approach (depending, of course, on the lighting conditions of the host planet) would be to instead use a photosynthetic layer. That would make the cyborg very literally green as well as semi-transparent too. It has the bonus of running completely silent as well as requiring no daily dose of chemicals, aside from those found in your typical life-supporting atmosphere.
It's also worth mentioning that the most efficient kind of photosynthetic layer would appear deep black as it's absorbing all of the light that lands on it.
The structure of such a layer would likely be nanograss - a nanotechnology technique which helps maximise the available surface area, boosting the amount of sunlight it can capture.
One of my personal favourite real-world concept designs featuring this kind of layer is the Nokia Morph - a concept phone which explores the ways in which nanotechnologies might be beneficial in the future.
How much power could it generate?
At the equator, the average sunlight received by the top of Earth's atmosphere is about 1.3kW/m^2, according to NASA. On average, it's 340W/m^2. About 48% reaches the surface, giving us an average of about 163W per square meter, accounting for clouds etc.
Meanwhile, the average human male's body surface area is about 1.9 square meters - most of this is shaded, but that wouldn't necessarily be the case for a well designed cyborg which doesn't have any need for clothing. Humans also consume about 97W of power, on average. Keep in mind that a large portion of this is used for heating and digestion too, which the cyborg has little need for.
So, assuming our cyborg has a male build, is in Earth-like conditions and has human-like power requirements, as well as a nanograss power generation surface, it could easily have 4 square meters of power generating capacity; in the absolute best case scenario with 100% efficient panels, that means the best it can generate is 4 * 163 = 652W, which is well above the 97W it needs, giving a possibly large enough margin for efficiency losses due to shade and non-ideal solar conversion efficiency.