I've got some different views on playing to the settings. Not things you'd do just to try to win normally, but optimizations towards this particular challenge.
1) A HOF game below Monarch means a warrior rush is usually best. The AI usually pops out a worker on the t18-19 interturn so you want to have two warriors there to attack. The odds of winning the 2v1 are OK (not great) but needs must, and this gets you 2 cities + 2 workers faster than anything else.
This is motivation for maximum AIs + No City Razing + Snaky Continents.
(N.B. For a space victory - on Pangaea or Terra - I would look for a much more extreme rush but Archipelago and a Hunting civ both make that more difficult.)
2) There's an art to rushing the AIs on the cheap. Even better if they "forgot" to research bronze working for Slavery, or Archery.
It's very definitely worth it if you can manage it since you get workers, religions, and sometimes granaries for free.
You wouldn't want to bull-rush an AI with 7 axemen though because that's quite expensive.
3) I'm not a huge fan of Oracle->Civil Service (although it is good practice if you're just trying to improve).
The game is a balance between very wide expansion, and teching to Sushi.
The main sources of research will be commerce tiles (in the very early game), internal overseas trade routes, coast tiles (FIN + colossus) and great people.
This means the priority wonders are:
- The Great Lighthouse (obviously)
- Colossus
- Everything else (probably Pyramids, MoM, Great Library and Oracle are on a similar power level).
Code of Laws is competing with Currency and Metal Casting (and maybe Alphabet), and you're spending 12 worker turns on Marble + chops. IMHO it's not worth it unless you can pick it up late.
4) I agree with a lot of the comments here about prioritizing Settlers and Workers over Granaries and Lighthouses until the initial wave of expansion is done - but the more successful your initial rush + expansion is, the faster this switchover comes (and the more wonders you can afford to build too).