The most common GC detectors I've run across are thermal conductivity (TCD), flame ionization (FID), electron capture (ECD), nitrogen phosphorus sulfur (NPS), and mass spec. Each has it's strengths and weaknesses.
A TCD just uses the carrier gas (usually helium) for detection. But it is relatively insensitive. The concentration of unknown must be relatively high.
An ECD typically uses radioactive sources. It is often ignored because of regulatory issues. But it works well.
Now for the FID. Think of it as a conductivity meter in a flame. You run your carrier gas into a burning flame (usually of hydrogen and air). The flame has two electrodes in it which measures conductivity. As a compound elutes it changes the conductivity. This is measured and sent to an intergrator. It's an oversimplified explanation, but it gives the general idea.
You can identify unknowns by retention time match, although that is not very specific. If you are using mass spec, you can get a better match through the spectrum. While you can estimate the concentration, most procedures require you to make a calibration curve. While a single point curve is possible, it usually is not well regarded by regulatory agencies. Multi point calibration curves are the normal rule.
Usually, you have an idea of what is present when you make and injection into the GC. This is especially true of detectors like FID. It is less true with an MS detector, which can also produce a spectrum match.