So I guess this gets back to my question of what is a 'complex' system- to run the sort of skill challenge based system you're describing, we'd need:
1) Scaling DC's- we had a table for this in 4e, could conceivably get the same thing for 5e.
2) Rules/guidelines for multiple success/failure when using skills- something like the Skill Challenge rules, hopefully with some explanation of how that looks in-game with social skills
But for low-level play we have rules for setting DC's already.
For 1, I think the flat math might handle things so as to make scaling unnecessary. This is more-or-less how Burning Wheel does it: reasonably flat math, but also adding a player metagame resource (Fate Points and similar tweaks) to enable players to pile on the bonuses if the maths gets bumpy.
No scaling DCs
and no Fate Points or similar might make players tend to shut down or turtle a bit - but I think you'd have to road test the maths to find out.
I think that 2 is perhaps the bigger deal.
Saying 'that's a big ask- that requires five successes before three failures' or 'that's easy- two successes before four failures' or something doesn't seem like anything that requires much more than a rough heuristic. So why can't we do this with the rules we have right now?
I think that if you want your pacing to come out nicely, you might want more thoroughly settled numbers - first worked out mathematically, than vigorously playtested. Notoriously, 4e took a long time to settle on numbers that worked both from the maths and from the "how does this game actually play at the table" point of view. (Escalating bonuses made this harder than it should have been - so D&Dnext starts more strongly on this score, I think.)
But the other thing that is needed - and I think it's more than an article, but not necessarily
mechanically very complicated - is advice to the GM on how to adjudicate. HeroQuest revised takes only a page or two to set out its core mechanic, but there is a lot of advice. Burning Wheel's mechanics are a bit more complicated, but still can be set out in 5 pages or so. But the advice is pretty detailed.
For me, a big issue with D&D is that its traditional way of putting pressure on the players is the risk of death. But given how trigger-happy D&D PCs are in any event, introducing this sort of pressure comes close to being guaranteed to turn your social encounter into a combat encounter. Burning Wheel has a lot of advice on how to frame and then adjudicate situations so that the players feel the pressure, but the stakes aren't "live or die". This is where I personally find the way the medusa is presented in Caves of Chaos a big disappointment. It seems like it's meant to be this intriguing situation full of potential - but how does a GM stop it turning into a blood bath almost from the get go?
This goes to scenario design - putting in stakes and hooks that aren't just about living and dying - but also monster design - giving monsters these things, so the GM can look up the Monster Manual and find non-live/die stakes and hooks built into the monsters. It also goes to PC build - letting the players build concerns, goals and hooks into their PCs that aren't just about living and dying (like the PC's honour that LostSoul played on in his paladin skill challenge). I think the background features would be a promising place to start for this sort of stuff (in 4e class and paragon path fill this space) but more development would be needed than what we currently have.
Anyway, this is some of the stuff that I would see a module covering - its not so much the resolution mechanics, which as you say can be fairly simple (although Duel of Wits gives us an example of more bells and whistles - it has social defence, for example), but (i) strong GMing advice, and (ii) ensuring that story elements - scenarios, monsters, PCs - work from the ground up to provide support for the sort of GMing that is required.
I hope this makes at least some sense!