You might be referring to this? That’s what I found from a quick search, at least.
If I understand correctly, this is a little different - from what I recall reading a few years ago, the speculation was that Netflix (and similar apps) rendered the content directly, bypassing the normal rendering stack. It would be the equivalent of, on a Linux system, bypassing the compositor (e.g. Mutter or KWin), and directly rendering the content (I believe SurfaceFlinger is the Android compositor). This means that when something like scrcpy uses the competitor API to capture the content, the content is literally not there, because it bypassed that system altogether.
By contrast, the secure flag just allows app developers to ask the OS to disallow screenshots, to prevent data leakage (e.g. of your banking details). It’s all rendered in the standard way, though.
This may not be accurate - it’s based on assumptions, and forum posts I read years ago, but it’s the best explanation I have right now. If anyone knows better, please feel free to correct me.
It’s not exactly what you’re looking for, and won’t be as seamless, but you might be able to leverage scrcpy.
It uses adb (you may need a fullfat distro for this - lineage may not support it), and allows you to view and control your Android device from a computer. It can also handle audio, and can be used wirelessly. The one caveat is protected content will probably not show up in the mirror - e.g. if you cast your screen and try to stream Netflix, it will likely be unable to send the Netflix video over. The last time I tested, it depended on the specific app, and which APIs they used under the hood (at the time, YouTube worked, Netflix did not).
The community forked the core a while back: https://github.com/openVoiceOS/
See also https://neon.ai/, which builds on top of OVOS, and is who the Mycroft company officially passed the reigns to.