I have implemented a "working" CMFB loop since my last post (https://open-source-silicon.slack.com/archives/C016HUV935L/p1669055666383189). In essence, I am feeding the output voltage into a op amp that acts as an error function. The output of the CMFB-focused-op-amp is then fed into the gate of the tail current transistor of the first stage OTA (of the actual op-amp itself). However, the input (low-pass) filter (that extracts the DC component of the output signal) to the feedback-op-amp is massive. Is there a good way to reduce the area of this filter (1000K, 2uF)? Thanks once again for the help and insights guys.

CMFB is needed for fully differential amplifiers, but yours has a single ended output (there is no output common mode to speak of, just the bias point of the output node). The output bias point is usually set by the feedback network around the amplifier. What does your overall circuit look like, i.e. how exactly is this amplifier used?

This amplifier would eventually used in a non-inverting or inverting configuration (closed loop, feedback). If I'm understanding you correctly, you're saying that if I have a feedback loop around this op amp (let's just saying a typical non-inverting set up), that I will not need CMFB since the feedback network will set my output bias point? If this is the case, should I simply set up a closed loop non-inverting around this op-amp right now and test it? Thanks for your insights @Boris Murmann

Yes, but it depends on what you are trying to test. One simple thing you can do is connect the amplifier in unity gain configuration and look at the step response. In this configuration you can also simulate loop gain for stability tests, see other posts.