I've tried similar approaches. From that opening line and with sane priors, you can probably get a pretty good idea of what the results were.

For me, and I suspect many others for whom all self-help and motivational techniques and hacks just "inexplicably" fail and which "they must be doing wrong", the problem is almost entirely within one single, simple assumption that seems to work naturally for the authors, but which is for me a massive amount of cognitive workload that is continuously taxing on my mental energy.

And said assumption that I refer to is precisely here:

A habit I'm working on developing is to ask a mental model of a Manager what I -should- be doing right now.

The question I shall ask, to illustrate my point, is: If you were programming a computer to do this (e.g. open a chat window with someone posing as a Manager for the appropriate discussion), how would you go about it?

More importantly, how does the program know when to open the window?

Suppose the program has access to your brain and can read what you're thinking, and also has access to a clock.

Well, there are three most obvious, simple answers, in order of code complexity:

1. Keep the chat window open all the time. This is obviously costly attention-wise (but not for the program), and the window is always in the way, and chances are that after a while you'll stop noticing that window and never click on it anymore, and it will lose all usefulness. It then becomes a flat tax on your mind that is rendered useless.
2. Open the chat window at specific intervals. This brings another question: how often? If it's too often, it gets annoying, and it opens too many times when not needed, and eventually that'll cause the same problems as solution 1. If it's not often enough, then you won't get much benefit from it whenever you would need it. And even if it is a good interval, you'll still sometimes open it when not needed, or not open it when it was needed more often that day or in the middle of an interval. We can do better.
3. Look for the kind of situations in which the Manager will help you, by reading what you're thinking about, and then whenever certain conditions are met (procrastinating, not doing any work, spending too much time reading wikipedia articles, etc.), bring up the chat window. However, this is a large endeavor, because the program has to be constantly running and reading every thought that passes by, and then using (read: computing, running) heuristics to tell whether the conditions are met (read: run a complex function with the current thoughts as arguments/parameters, for every single given thought).

See, while I was writing this, I had forgotten about a specific work-related thing I was supposed to do at a certain condition. It's only when I wrote point 3 above that my brain actually connected this to "checking for events", which led to "I have events to check for!" which led to "Oh, right, that person got back, I should go ask them X".

The key point being that the very thought of even checking for conditions upon which to act is something that does not occur naturally or on its own for me - it has to come about by being linked to from another thought and brought to my conscious attention. Any technique that relies on consciously doing X inevitably stumbles on this key factor for me.

Running computations on every single thought all the time is extremely tiring and mentally exhausting. It's much more daunting than any task I would usually need "motivation" for. It means I stop after every few thoughts and think of the thing I have to remember to do. And then remember to think that I have to think about this again in a few more thoughts. And then try to resume whatever other thoughts I had. It's pretty much impossible to focus and concentrate on anything while doing this.

Which means whenever the set of conditions for talking to the Manager are met, I will not automatically open the chat window. It just won't detect the conditions. The conditions won't, on their own, open the chat window - the conditions themselves (I'm tabsploding on wikipedia) were not designed such that they always open the chat window with the Manager each time they happen.

So the tabsploding process happens, without ever calling on the remote parts of my brain that have little bits of code to open chat windows when tabsplosions happen, and so those remote parts of my brain keep on sleeping, and so chat windows do not open, and so tabsplosions go on merrily uninterrupted for hours until I read an article about business management, and the word management triggers me to remember the Manager process, and then I suddenly realize that I've been procrastinating all this time and need to get back to work (Note: I get back to work without even needing said Manager chat window, by this point, so the problem is clearly not "motivation" in this case).

And all that is the hidden assumption, the obvious thing that no one mentions in "making a habit of doing X" or "using GTD" or "using pomodoro". It's the single most brain-computationally-intensive process I can think of that people have ever actually seriously implied I should use. My subconscious, unfortunately, doesn't do it for me. It seems like most other people have it easier. Well, good for them. I'm still stuck here unable to realize that I need to do the dishes, and so I keep on reading forums, and my forum-reading thoughts don't have any bits dedicated to remembering whether or not dishes need to be done, so the forum-reading begets more forum-reading and tabsploding, and my mind never brings up the issue of having something to do.

And yes, this applies to meta concerns. So training myself to be more mindful and conscientious of these things fails because I fail to think of applying techniques to make myself more mindful and conscientious. Everything I've tried has failed to produce the amazing results others report.

I have no idea of how common this problem is, or whether nootropics might be a solution.