1. CoDzilla. The power level of the spell casters was kind of off the charts in 3.5. Pathfinder kind of helped this by buffing the martials and fixing the worst offenders. My solution would be to lower spell DCs and buff saving throws so at higher levels you only have a small% chance of flunking a save vs a save or suck or a save or die. This is in effect similar to how it used to work in AD&D. Spell DCs would be 10+ the level of the spell (capped at 20) and some feats could increase this.
I came at this the other way. Spell DCs are 10+ability bonus. The save DC does not depend on the level of the spell. This accomplishes the same purpose in my opinion - allowing saving throw bonuses to increase more rapidly than the target DCs.
Additionally, Clerics are given a list of known spells by level, similar to Sorcerers. Only wizards are allowed to know all spells, and this is generally controllable by restricting spell ability and the general difficulties with owning spell books.
Druids would also be overhauled and lose the animal companions going back to something like the 2nd ed Druid and the feat natural spell would go bye bye. The spell list could also be cut down eliminating or rewriting the problem spells.
Druids went bye bye completely under my rules, for a lot of reasons, and yes so did the feat 'Natural Spell'. I replaced the concept of the Druid with the more generic and more balanced Shaman class from Green Ronin and reshuffled some spells. You can still make a Druid in every sense in my game, but it will be significantly less potent in many ways than the stock 3.X Druid.
Specifically scaling buff spells like divine favor and divine power. AD&D had buff spells and Prayer for example was actually a decent spell. Spells like Greater Magic Weapon, Divine Favor/Might either need to go or become a static bonus. Overall I would be trying to drag the numbers porn down more towards AD&D/BECMI levels.
I haven't had too big of a problem with these things. I find buffs to be somewhat self-limiting. The main buff spells that are problematic IMO are the shape-change spells. Those are the ones that need redefining to more predictable bonuses and caps. Otherwise, just be careful with things like Persistant spell.
I generally have no problems with martial characters ability to deal damage. The only thing you mention I even have a problem with is Rapid Shot, and that's simple because its a no brainer that ups the number of attack roles and the time required to resolve attacks while simultaneously advantaging the best response to any tactical problem - hit it from range. Martial characters being primary DPS isn't a problem I've got. My biggest problem is the relative value of a martial character outside of combat compared to a spellcaster, which is where martial character's need buffs and problem solving capability.
I have no issue with ability caps. Maintain some sanity in your item distribution and availability of magic, and things should be fine.
Overhaul the skill system and even out the spread between the classes for skills. Most classes would have 4-6 skill points with wizards getting 2 and rogues getting 8. DCs would top out at about 25 or 30 though. Pathfinder has the best skill system IMHO our of 3.0,3.5, 4th ed and SWSE but it could do with some tweaks.
Under my current system, which has slightly more skills than standard 3.X, spellcasters generally get 3 skill points per level, where as martial classes vary from 4 (in the case of a fighter) to 11 (in the case of a rogue). I've made a variety of other tweaks that make skills more epic sooner, but the main things is to allow all martial classes to be something of skill monkey's relatively easier and to reduce the ability of spells to substitute for skills to an absolute degree.
Overhaul multiple attacks. instead of having 3 attacks at +11/+6/+1 I would just allow 3 attacks at +11 and eliminate the full attack. Let fighter types move their full movement rate and unleash. This is similar to how it worked in AD&D and the fighter was actually nerfed in this regard from 2nd ed to 3.0.
I've had no problems with the current iterative attack system beyond the math complexity. The advantage of 3.X standard iterative attacks is that they create much smoother damage curves. You seem to be busy buffing fighters in areas that they really don't need to be buffed in.
Drop wealth by level and the ability to purchase magic items.
Yes.
Go back to 2nd ed item creation rules.
No, though there does need to be some greater attention paid to balance and some general rules overhaul.
Yes. I generally just made it impossible to create divine wands.
8. Eliminate X2 and X3 damage on crits. Replace with max damage or an extra dice of damage. Reduce the rocket tag effect and the importance of weapons with 18-20 threat ranges.
I've got no problem with crits as written. Again, the problem with game balance is not that fighters can produce lots of damage. That's ok, and easily balanced against. The problem is that spellcasters too easily dominate the action economy (the same could be said of some monsters, requiring too much dependency on absolute immunities at high level), and martial classes are too easily overshadowed in problem solving ability out of combat.
9. Tweak the classes and monsters to account for the reduced damage and spell effects. Perhaps return weapon specialization to its AD&D levels of glory.
No. Weapon specialization is and was always a bad idea that hurt the game, particularly in its AD&D incarnation. I've kept it in the game because players expect it, but taken steps to mitigate its bad side effects - like being disappointed to find weapons other than that one weapon type you use well.
You might search for 'Playing like Celebrim', where I do some of my class write ups and describe my general philosophy.