In the Eclipse campaign, I play a human who has been heavily modified with cybernetics. I've been playing this character since the start of the campaign, so I've had a lot of time to think about theoretical alternate implementations of cybernetics. I want to be very clear that this is not intended as a critique of Eclipse's rules - this is a thought experiment and nothing more, based around fusing two rules concepts that have worked out pretty well so far in Dust to Dust.
The quick summary of the structure is a mana-based system with a variable spell list based on opportunity cost. I like using a mana-based (well, obviously it's actually Bio-Energy) system because when I imagine the cybernetic mechanics as a whole, I envision a single internal energy source rather than one battery per module. Things can be justified in-play either way - that's just what makes the most sense to me, especially given that the reason you have to spend CP (as a stand-in for time and effort) to add new modules, other than game balance, is that you're increasing your personal reserves of Bio-Energy. I like the idea that I could re-route power from one module to another if I'm not going to have a chance to use something, or it seems lower-priority. With characters in general, I like having that feeling of adaptability and fluidity; it's what I loved about druids in both WoW and D&D 4e.
As for the variable spell list and opportunity cost, what I'm imagining is that your modules form your "spell list," but you can just go get a cyberdoc to change out your modules if what you've got isn't what you need. I think that one possible criticism of this approach is that one cyborg can be any cyborg - the sense of personal differentiation is lost - but I would counter that by pointing out that we haven't worried too much about differentiating one spellcaster of a given type from another spellcaster of the same type, at least not in terms of CP expenditure. There are also other ways to introduce personal differentiation, which I'll get into in a bit.
In exchange for some loss of differentiation, this structure offers a thematic expression of transhumanism, as the body becomes an infinitely adaptable tool with which to solve problems; while Eclipse's current system does include this theme, I think it is taken to a new height if modification to solve one problem doesn't lock the character into that kind of modification for the rest of his life. It also intensifies the contrast between mutation (involuntary genetic adaptation to suit one environment or set of challenges) and cybernetics (voluntary scientific adaptation to suit any environment), and incidentally also psionics (voluntary mental and mystical adaptation to surpass limitation). Now, I'm hardly recommending that cybernetics should be inherently better than the other two paths; as it stands, mutation reaches heights ofadaptation to niches that cybernetics does not.
The systemic limitation of this idea, as in the existing system for cybernetics, is connected to "slots" of the body, each of which can be upgraded from Consumer-grade to Commercial-grade to Military-grade. These can be reconfigured, of course, to possibly narrow down the number of "spells" a character can access at a time. It is an open question, to me, as to whether I'd want to require a character to have the prereqs for something installed in order to use it - that is, would the Skin 4 module require that you also had the Skin 2 module? It could really go either way, but if it doesn't, the game's balance becomes much more delicate, simply because some of the progressions are at their best at Military-grade, while some others are good enough for 99% of all cases at Commercial-grade.
I also want to go over some of the apparent problems with this approach. The first is that it's very difficult to assign mana costs to some of these abilities in a coherent way - for example, we've never been faced with the situation posed by the cybernetic Stealth progression of a night-only Meld, night-only Invisibility, and anytime Meld in a mana-based system. How would you price an anytime Meld, when it's sometimes the same effect as a night-only Meld? In the existing system based on purchasing per-day usage, this is a non-issue; anytime-Melds are useful at night when you've otherwise exhausted your daily usage of night-only Melds. The immediate answer is to say that you can now activate Melds during the day, presumably at the same Bio-Energy cost as you had been using for your night-only Melds. Is that enough, though? If you can just go get a cyber-doc to replace it with something else at nightfall, it becomes a lot less desirable. For that matter, the only reason you still want a night-only Meld once you have a night-only Invisibility is that the mana cost of the former is presumably much lower - and perhaps you would also slap a small casting time on Invisibility, to leave Melds as the ideal panic-button effect.
My preferred answer here is to give the anytime-Meld module a secondary effect, perhaps modifying the mana cost of any Melds you cast on yourself at night, or allowing you to extend your Meld effect to another character at night, as long as that character remains in near-contact with you. Even a minor situational benefit is probably enough of a lure to keep me from feeling like it's a waste. Come to think of it, though, night-only abilities are either the selling point or the bane of this system, depending on whether you think it's okay for every cyborg to go under the knife twice a day. If you regard this as a problem, I'd suggest that such modules could still function during the day, perhaps at a greatly heightened mana cost.
Another problem to tackle is the always-active effect - really anything that the current system sells at a fixed CP cost rather than a per-day cost, such as claws, Security Countermeasures, or secret compartments. The clear answer here is to assign a mana total that is "tied off" in sustaining the module, presumably priced according to perceived utility - it'd be tough to argue that any grade of Security Countermeasures is worth as much mana as Short Claws. Such prices are difficult to assign, but the same could be said of CP costs for the abilities in the current system, so I don't regard it as a weakness of the design.
Miscellaneous modules, as "slotless" modules, are just as "free" here as they are in the current design, but giving each of them a Bio-Energy cost as suggested above sets a soft cap on their usage. If this came to grant too many desirable passive abilities, I'd recommend including a hard cap by fixing a number of Misc Module slots that each grade of neural interface can support. In the existing system, cybernetics is the king of passive abilities, especially at Military-grade.
I'm envisioning a system in which a mission that requires environmental adaptation is open to all cyborg characters, as long as they're willing to temporarily dump their existing full-conversion modules. It's another way to make sure that no one on the module is using cybernetic stealth, without requiring NPCs with True Sight to be part of the mission. Because they are getting some other kind of functionality out of their cybernetics, I think the cyborgs who would normally use stealth modules will feel less like the game has solved for their build. After all, they spent CP on Bio-Energy, not on their stealth modules.
If one wanted to add personal differentiation back into the system, there are a few different ways to handle it. Some modules can be permanent installs, such as archaeotech, or even black-market modules that have a few corrupted files. For that matter, "one module is no longer removable" would be a valid flaw to pull from a Deck of Possibilities. Finally, the cyborg's neural interface is the one thing that can never be removed, only upgraded; Commercial-grade and Military-grade neural interfaces could include branching options with various strengths. (This last idea somewhat parallels the Ebon Aspect's sixth- and tenth-mystery abilities in DtD.)
Anyway, these are just some ideas I had - difference for the sake of difference, trading one set of design strengths for another. I hope you've enjoyed my idle brainstorming. I am amused to note that in the time it took me to write this, Kainenchen started playing with her Valentine's Day present: Deus Ex: Human Revolution.