Surrealone

Per 6e1 p193:
Dispelling Incoming Attacks
A character can use Dispel to protect himself from incoming attacks, but he must have a Held Action. Assuming the Dispel applies to the attack, he uses his Held Action to “attack” the incoming attack with his Dispel. He rolls his Dispel dice (he doesn’t have to make an Attack Roll, though this does constitute an Attack Action). He Dispels the attack if the total of the Dispel dice exceeds the Active Points in the attack. The Power Defense of the character making the incoming attack (if any) doesn’t apply to reduce the Dispel effect roll.
 
Thus, to answer the original question:
No, you cannot Abort (key word: abort) to Dispel per RAW … and I would not permit it.  Key to this is that it requires an attack roll and, thus, is an Attack Action (i.e. not a defensive action).  This is why a Held Action is required …

Per 6e1 p193:
Dispelling Incoming Attacks
A character can use Dispel to protect himself from incoming attacks, but he must have a Held Action. Assuming the Dispel applies to the attack, he uses his Held Action to “attack” the incoming attack with his Dispel. He rolls his Dispel dice (he doesn’t have to make an Attack Roll, though this does constitute an Attack Action). He Dispels the attack if the total of the Dispel dice exceeds the Active Points in the attack. The Power Defense of the character making the incoming attack (if any) doesn’t apply to reduce the Dispel effect roll.
 
Thus, to answer the original question:
No, you cannot Abort (key word: abort) to Dispel per RAW … and I would not permit it.  Key to this is that it requires an attack roll and, thus, is an Attack Action (i.e. not a defensive action).  This is why a Held Action is required …

My usual route for someone who wants to be able to do everything is to say yes but your "theme" is therefore versatility and because you can do everything you can do it less well than specialists, so if the usual energy blast is 10D6, every power man has 8d6 or 9d6.  If the usual force field is 20PD 20ED thenevery power man has 17 or 18.
 
everything, just a little bit less of it....
 
Doc

Don't try to build it into the gun.  Instead, have the character make a Concealment roll when placing the weapon on his/her body and use the result of that roll (in contested fashion, if desired) to determine how concealed the weapon is.
 
I suggest this approach because strapping a holdout piece to one's ankle, concealing a firearm in an IWB holster, and slinging an Uzi or shotgun under one's coat -- and moving about with such things equipped … all have one thing in common: a degree of skill when it comes to clothing choices, precise position on the body, and movement with the concealed weapon in place.
 
You might also consider a complementary PS: Concealed Carry.  I suggest this because practically every person who carries concealed … has a bin of holsters s/he has tried in various combinations with clothing until he/she has arrived at what works for his/her body type, clothing preferences, and movement tendencies.  I also suggest this because someone who is carrying a firearm for the very first time tends to move much less naturally than someone who has been doing it for years.  Movement characteristics in conjunction with clothing and body type  are important because certain motions can cause a concealed firearm to 'print' (i.e. have the outline of it shown on a cover garment).
Example -- an ankle holster with a small firearm may print if someone is wearing loose slacks and the wind blows hard directly against the side of the leg where the firearm is located; someone who is accustomed to carrying concealed will almost instinctively move to positionthe ankle holster and firearm in the lee of the leg relative to the wind, thereby avoiding such printing. Example: an IWB (inside waistband) holster with a firearm located at the 4 o'clock position beneath a tucked-in shirt is prone to printing on that shirt when someone stretches to reach something high, specifically due to the tension applied to the otherwise loose shirt being applied directly over the firearm; someone who is accustomed to carrying concealed will naturally tend to use the left arm/hand to perform that reach, thereby tightening/tensioning the material on the 8 o'clock position of the body rather than the 4 o'clock side  
A Concealment roll each time a weapon is strapped on also makes sense … because something as slight as the difference between the 3:30 and 4 o'clock positions can make a substantial difference in print levels or noticeability.  Colour choices also matter, and as another example, consider that open carry of a black firearm in a black/grey custom leather holster … while wearing grey shorts and a black shirt … with black socks and black/grey shoes … can yield an openly-carried (i.e. in plain sight) firearm that next to no one notices simply because it blends into the colour palette of the clothing being worn with a rudimentary camouflage effect.  This is basically what belt buckle firearms rely on -- a very limited form of camouflage that causes people to overlook them.  In reality that form of camouflage isn't something built into the belt buckle gun. Instead, it's something specific to the relationship of the gun, the belt buckle holster on which it's carried, the clothing choices of the person wearing it to 'gel' with that belt and belt buckle, the body type of someone able to wear a belt buckle that large, and the human tendency to overlook certain things yet perk up and pay attention to others while people busily go about their lives.  Concealment is likely the skill that best suits someone's ability to identify and effectively leverage that relationship.

Don't try to build it into the gun.  Instead, have the character make a Concealment roll when placing the weapon on his/her body and use the result of that roll (in contested fashion, if desired) to determine how concealed the weapon is.
 
I suggest this approach because strapping a holdout piece to one's ankle, concealing a firearm in an IWB holster, and slinging an Uzi or shotgun under one's coat -- and moving about with such things equipped … all have one thing in common: a degree of skill when it comes to clothing choices, precise position on the body, and movement with the concealed weapon in place.
 
You might also consider a complementary PS: Concealed Carry.  I suggest this because practically every person who carries concealed … has a bin of holsters s/he has tried in various combinations with clothing until he/she has arrived at what works for his/her body type, clothing preferences, and movement tendencies.  I also suggest this because someone who is carrying a firearm for the very first time tends to move much less naturally than someone who has been doing it for years.  Movement characteristics in conjunction with clothing and body type  are important because certain motions can cause a concealed firearm to 'print' (i.e. have the outline of it shown on a cover garment).
Example -- an ankle holster with a small firearm may print if someone is wearing loose slacks and the wind blows hard directly against the side of the leg where the firearm is located; someone who is accustomed to carrying concealed will almost instinctively move to positionthe ankle holster and firearm in the lee of the leg relative to the wind, thereby avoiding such printing. Example: an IWB (inside waistband) holster with a firearm located at the 4 o'clock position beneath a tucked-in shirt is prone to printing on that shirt when someone stretches to reach something high, specifically due to the tension applied to the otherwise loose shirt being applied directly over the firearm; someone who is accustomed to carrying concealed will naturally tend to use the left arm/hand to perform that reach, thereby tightening/tensioning the material on the 8 o'clock position of the body rather than the 4 o'clock side  
A Concealment roll each time a weapon is strapped on also makes sense … because something as slight as the difference between the 3:30 and 4 o'clock positions can make a substantial difference in print levels or noticeability.  Colour choices also matter, and as another example, consider that open carry of a black firearm in a black/grey custom leather holster … while wearing grey shorts and a black shirt … with black socks and black/grey shoes … can yield an openly-carried (i.e. in plain sight) firearm that next to no one notices simply because it blends into the colour palette of the clothing being worn with a rudimentary camouflage effect.  This is basically what belt buckle firearms rely on -- a very limited form of camouflage that causes people to overlook them.  In reality that form of camouflage isn't something built into the belt buckle gun. Instead, it's something specific to the relationship of the gun, the belt buckle holster on which it's carried, the clothing choices of the person wearing it to 'gel' with that belt and belt buckle, the body type of someone able to wear a belt buckle that large, and the human tendency to overlook certain things yet perk up and pay attention to others while people busily go about their lives.  Concealment is likely the skill that best suits someone's ability to identify and effectively leverage that relationship.

Regarding a minimum cost to be able to apply limitations to differing types of skill levels: as a reminder, the application of limitations to the various types of skill levels is a GM call in 6e -- whether we're talking about Combat Skill Levels or plain old Skill Levels (the latter of which would be relevant for a bonus to the Concealment Roll).
 
Per 6e1 p71:
COMBAT SKILL LEVELS WITH LIMITATIONS
With the GM’s permission, characters can put Limitations on CSLs (for example, to build equipment, like a laser sight for a gun). The GM may restrict which types of CSLs a character can Limit; for example he might rule that only 3-point or more expensive CSLs can have Limitations. Unless the GM rules otherwise, CSLs with Limitations can only increase the user’s OCV, not DCV or damage.
 
Per 6e1 p88:
With the GM’s permission, characters can put Limitations on Skill Levels. The GM may restrict which types of Skill Levels a character can Limit; for example he might rule that only 3-point Skill Levels can have Limitations.

Don't try to build it into the gun.  Instead, have the character make a Concealment roll when placing the weapon on his/her body and use the result of that roll (in contested fashion, if desired) to determine how concealed the weapon is.
 
I suggest this approach because strapping a holdout piece to one's ankle, concealing a firearm in an IWB holster, and slinging an Uzi or shotgun under one's coat -- and moving about with such things equipped … all have one thing in common: a degree of skill when it comes to clothing choices, precise position on the body, and movement with the concealed weapon in place.
 
You might also consider a complementary PS: Concealed Carry.  I suggest this because practically every person who carries concealed … has a bin of holsters s/he has tried in various combinations with clothing until he/she has arrived at what works for his/her body type, clothing preferences, and movement tendencies.  I also suggest this because someone who is carrying a firearm for the very first time tends to move much less naturally than someone who has been doing it for years.  Movement characteristics in conjunction with clothing and body type  are important because certain motions can cause a concealed firearm to 'print' (i.e. have the outline of it shown on a cover garment).
Example -- an ankle holster with a small firearm may print if someone is wearing loose slacks and the wind blows hard directly against the side of the leg where the firearm is located; someone who is accustomed to carrying concealed will almost instinctively move to positionthe ankle holster and firearm in the lee of the leg relative to the wind, thereby avoiding such printing. Example: an IWB (inside waistband) holster with a firearm located at the 4 o'clock position beneath a tucked-in shirt is prone to printing on that shirt when someone stretches to reach something high, specifically due to the tension applied to the otherwise loose shirt being applied directly over the firearm; someone who is accustomed to carrying concealed will naturally tend to use the left arm/hand to perform that reach, thereby tightening/tensioning the material on the 8 o'clock position of the body rather than the 4 o'clock side  
A Concealment roll each time a weapon is strapped on also makes sense … because something as slight as the difference between the 3:30 and 4 o'clock positions can make a substantial difference in print levels or noticeability.  Colour choices also matter, and as another example, consider that open carry of a black firearm in a black/grey custom leather holster … while wearing grey shorts and a black shirt … with black socks and black/grey shoes … can yield an openly-carried (i.e. in plain sight) firearm that next to no one notices simply because it blends into the colour palette of the clothing being worn with a rudimentary camouflage effect.  This is basically what belt buckle firearms rely on -- a very limited form of camouflage that causes people to overlook them.  In reality that form of camouflage isn't something built into the belt buckle gun. Instead, it's something specific to the relationship of the gun, the belt buckle holster on which it's carried, the clothing choices of the person wearing it to 'gel' with that belt and belt buckle, the body type of someone able to wear a belt buckle that large, and the human tendency to overlook certain things yet perk up and pay attention to others while people busily go about their lives.  Concealment is likely the skill that best suits someone's ability to identify and effectively leverage that relationship.

Because when you pull and fire a gun, it's hardly 'inobvious'... and an IAF remains inobvious even during use.  A great parallel example would be a cell-phone vs. Bluetooth-enabled Airpods.  The cell phone would be an OAF (much like a gun) because it's inobvious when vibrating in your pocket but quite obvious and quite accessible when it's in hand and being used ….  while the Bluetooth-enabled Airpods would be an IAF because they remain inobvious throughout use (unlike a gun).
 
More important, I think, is the fact that concealing a weapon on one's body is absolutely a skill … and is an example of why we have Concealment in the book.

Why is it not as simple as just using what's in the book to properly simulate what actually happens when someone, you know, conceals (hint: skill roll) a gun on his/her person … rather than taking some hackneyed approach with an IAF that is actually obvious while being used … and handwaving when you need not do so?

It's been a long time since I've seen that -- largely because a) most of the people I've gamed with are mature enough not to take a value for a limitation that doesn't limit them and b) most of the GM's I've gamed with are seasoned enough to spot that sort of thing a mile away and stomp it out on those rare occasions where it comes up.  (Note: Both a) and b) result in no need to pay back the points because no cost break was received in the first place.)

I play both heroic and super-heroic games.  When gaming at the heroic level, equipment tends to matter and guns are quite lethal. However, even at the super-heroic level, it is largely the advantages and limitations that define a power. 
 
Let's boil it down to brass tacks.  If you do this for limitations then it should also be done for advantages -- because the same logic applies to both, and the approach should be consistent for both.  And since you're all about super-heroic play, just try making a Mental Entangle in 6e without a list of advantages/limitations if you have any doubts on this topic.  And note: that's just to make the basic Mental Entangle -- onto which someone then might need to lump additional advantages (IPE, for example) and limitations (charges, only works when X, etc.) to take that Mental Entangle and make it fit properly into the overall character concept. 
 
In fact, Mental Entangles are great examples of why having only 1 or 2 advantages/limitations (or even a blanket cap on max number or value of advantages/limitations) just doesn't make much sense to me : with only 1-2 of either, the Entangle power could no longer be properly defined into a Mental Entangle AND something with advantages/limitations that gel it into the character's concept...
 
P.S. Sure, I get that this topic started ONLY about limitations, but again, the logic applied to dumbing down limitation lists also applies to advantage lists, and I firmly believe that if you run with dumbing things down, you need to do it to not only what affects the divisor of a power's costs, but to that which affects the multiplier of a power's costs, too -- just to be consistent in how you dumb down the math. 

No, you can't really get by with just one limitation while maintaining the level of granularity Hero offers.  As an example of why, let's use your example.  Would your Real World limitation preclude a common firearm from firing in space?  If your answer is 'yes' then I have to call BS, because modern ammunition contains oxidizers that allow it to fire under water, in space, etc.  And there, sir, exists an example of the pitfalls of vague, non-granular limitations (i.e. one person thinks it means one thing, while a more or differently-informed person thinks it means another)...
 
Thus, if you want that 'Does Not Work In Space' limitation, you should receive it … and then we're back to the GM having to rule as to whether that's worth any points (since, you know, his game might not involve going into space).
 
Again, the GM is the existing solution to whether limitations are appropriate to the campaign setting … based on whether they will or won't come into play (and, if so, how often).  And there's nothing wrong with granular lists of advantages or limitations if they serve to define a power properly.  After all, the entire point of eliminating some powers (from previous versions ... like say, Force Wall) and then using advantages/limitations to achieve the same result with a more generic power (a la Barrier) … was to simplify the toolkit without any loss of granularity.  What you propose … is to chuck a large amount of granularity out the window.  A GM is certainly welcome to do that, but a lot of folks play this game precisely because it offers a level of granularity not seen in most RPGs.

This proposal feels like a solution in search of a problem.  I say that because given the presence of a GM who has final adjudication rights regarding his/her game, there just isn't a problem to be solved.  Now if we lacked GMs, then we'd be missing a solution to a number of imperfections in the game system, but as it stands, today, the GM -is- the solution to certain issues.
 
Because it's not broken, I don't think there's anything to fix/address on this front...

This appears to be Desolidification with the Cannot Pass Through Solid Objects (-½) limitation and the Does Not Protect Against Damage (-1) limitation.  Given that there's already a well-established way to build the effect of the proposed new power, I see the proposed new power as wholly un-necessary.
 
That said, I think the Barrier Chart you've provided is an excellent additional to optional rules to help govern what someone (who builds out the aforementioned Desolidification w/ its limitations) can move through -- for GMs who desire a pre-baked chart of such things.
 

Out of curiosity I took at look at both GRAB and STRETCHING, as I seemed to recall rules about moving a grabbed character in there somewhere.  Here is the gem I was looking for.
 
Per 6e1 p286:
A character with Stretching can reach out, Grab a character or object that’s within the range of his Stretching, and pull it to him (assuming he has the STR to move it). (He cannot pull himself to the Grabbed character or object, he drags it to himself.) If the target is an object, this typically takes a Full Phase Action. It only requires a Half Phase Action if the total of (meters Stretched + meters the character has to pull the object) is less than or equal to half his Stretching and the object is non-resisting. If the target is a person or the like, the GM may rule that making the Grab ends the Stretching character’s Phase. However, since characters can Grab and Squeeze, Slam, or Throw, most GMs allow characters with Stretching to Grab and Drag To Myself as part of one Attack Action. The GM may allow a Grabbed character to have an Action that takes no time to brace himself or use STR to resist being dragged.
 
The takeaway, here, should be that triggered UAA movement should not be needed so long as the attacker has enough STR to move the target to him/her (or enough STR to move him/herself to the target if allowed by the GM).  IMHO, this bit of RAW also strongly suggests Stretching was likely intended to be able to model what Spidey does...
 
As an aside, if the GM requires the Grab & Pull to take more than the Full (or potentially Half) Phase, then you're back to triggered UAA to build what you desire, but I can't imagine a GM requiring that given how cinematic and appropriate the single Grab & Pull activity tends to be.
 
 
 

I believe the SFX of webbing a target and pulling it to the character can be readily boiled down to a focus-based grab maneuver with a triggered UAA movement power (that's triggered by a successful grab using the focus).  Thus, I think your stretching option makes the most sense, here.

I believe the SFX of webbing a target and pulling it to the character can be readily boiled down to a focus-based grab maneuver with a triggered UAA movement power (that's triggered by a successful grab using the focus).  Thus, I think your stretching option makes the most sense, here.

So female characters who carry purses in your games have to pay points for the retention/shoulder straps, then (lest there be a double-standard).  Got it.
 
And yes, I would consider that line of thinking too onerous to niggle with in a superheroic game.  But if that's how you like it in your game, that's wonderful for you!

A side issue is actually specifically mentioned...6E1 p. 303.  Teleport, UOO:
 
 
A rifle's got at least a -2 to hit;  a grenade's probably -6.  (Off the top of my head w/o cross referencing.)  This is a potential issue regardless of the actual attack definition, if it's targeted.  I'm not a fan of the AoE notion, saying oh I can affect everything all at once.....no.  You're asking for a form of multi-attack AND trying to  get a limit on it by claiming it's only vs. foci.
 
With teleport...the effect more or less fits, and whole objects are being teleported so it's legal...but it's pretty cheesy.  Even if there's a campaign rule saying minimum Teleport 5m, the baseline's only 5 points.  Even HEAVY advantages won't make it very expensive and it's basically unresistable.  
 
I agree with OP's assessment on the drain/suppress.  Doesn't act correctly.

As I see it, the trouble with your proposed approach is that it's all or nothing ... meaning it either works or it doesn't against all Focii … whereas you indicated what you wanted was something Lucky Lass was 'hoping to cause him to drop his rifle and have his belt and combat harness fall off him'.  To me, your 'hoping to' qualifier was read as indicating the power should NOT be all/nothing … and that there should be a chance for an unlucky event to happen to each of the Focii, rather than it affecting all of them ... or none of them.

Thus, instead of an AoE I would expect this to look something more like a Multiple Attack against each Focus (taking the penalty for targeting a Focus, of course … and the Multiple Attack penalty) … and using an appropriate power (TK? Martial Disarm? Whatever fits your special effects, really...) against each Focus you wanted to Disarm from the target.
 
I definitely don't see Only to Disarm as -1 on the TK; more like -1/2 … but that's really up to your GM, I suppose.

Funny!  I say that because my idea of the perfect RPG system entails exactly one book ... of 400 pages or less... resting on a coffee table with no other materials around it.  That's actually -why- I was drawn to Champions: 1st through 3rd Editions were tiny and 4th Edition ended up being exactly what I was looking for, as one book absolutely could do it all.
 
I think it got kind of ridiculous with 5th Edition (too big) .... and the 2-volume 6th set is absurd in its verbosity ... so colour me pleased to see Champions Complete, since it was a return to what I felt Hero System should be.  You just shouldn't need a crate of books or a directory full of PDF's to play any RPG, IMHO.

Funny!  I say that because my idea of the perfect RPG system entails exactly one book ... of 400 pages or less... resting on a coffee table with no other materials around it.  That's actually -why- I was drawn to Champions: 1st through 3rd Editions were tiny and 4th Edition ended up being exactly what I was looking for, as one book absolutely could do it all.
 
I think it got kind of ridiculous with 5th Edition (too big) .... and the 2-volume 6th set is absurd in its verbosity ... so colour me pleased to see Champions Complete, since it was a return to what I felt Hero System should be.  You just shouldn't need a crate of books or a directory full of PDF's to play any RPG, IMHO.

Toons seem to take body, they just don't die from it.

 
Except of course according to Who Framed Roger Rabbit where they have the Dip.
“Remember how we always thought there wasn't a way to kill a Toon? Well, Doom found a way: turpentine, acetone, benzene. He calls it the Dip!”
―Lt. Santino

Appropriate amounts of per-turn Regeneration (with the ability to resurrect and heal limbs) fail to properly represent this how, exactly?  I ask because the ability to be Toon-like 'bonked' is effectively a special effect of the ability to die and recover from it.  That's Regeneration in Hero … and a LOT of it at a very high cost if it's supposed to be near-instantaneous.  Do note that even Wile E. Coyote peels himself off of rocks or walks away looking like an accordian after being pancaked … taking roughly a Turn (12s) to do it as he exits a scene in an amusing way … so I don't think you need to recover faster than that.
 
You probably need a high REC to go with it so that you have some END and STUN, of course … unless you forsee the character needing to take it easy like Wile E. Coyote does (when he goes back to the drawing board) before it goes back into action.
 
Put another way, I think we already have a power and appropriate characteristics that do exactly what you want. Thus, I don't see a need for another one.  It probably costs more than you want it to cost, but it's so stupidly powerful that it SHOULD cost a ton, IMHO.  Even your two-stage example of inability to lose limbs is actually covered by Regeneration; all you need to do is use the optional Hit Location rules to determine which limbs were hit, and consider them disabled or destroyed if enough BODY is done … then let Regeneration with Limb Healing ability do its thing to bring them back.  No special tables or limitations needed, at all … and no need to create a custom power around it; just use the optional rules you already have around Hit Locations in conjunction with Regen and you're golden.
 
That said, if you insist that this isn't a special effect of Regeneration and are absolutely intent on building some janky house-rule power around it, then I believe that the cost to achieve what you want using Regeneration and Recovery should be your guide as to the appropriate cost of whatever you build.  i.e. I think this should cost far more than the Automaton powers tend to cost if it's to be near-instantaneous.  To put that into perspective, Recovering 20 BODY per turn with the ability to Heal Limbs and Resurrect costs … (wait for it) … 345 active points.  So that's the kind of power and pricing you're talking about when you want someone to be unkillable and be able to get up and walk away from Wile E. Coyote-like deaths in Hero in 12 seconds.  (And that doesn't include the cost of a high REC to go with it...)

Appropriate amounts of per-turn Regeneration (with the ability to resurrect and heal limbs) fail to properly represent this how, exactly?  I ask because the ability to be Toon-like 'bonked' is effectively a special effect of the ability to die and recover from it.  That's Regeneration in Hero … and a LOT of it at a very high cost if it's supposed to be near-instantaneous.  Do note that even Wile E. Coyote peels himself off of rocks or walks away looking like an accordian after being pancaked … taking roughly a Turn (12s) to do it as he exits a scene in an amusing way … so I don't think you need to recover faster than that.
 
You probably need a high REC to go with it so that you have some END and STUN, of course … unless you forsee the character needing to take it easy like Wile E. Coyote does (when he goes back to the drawing board) before it goes back into action.
 
Put another way, I think we already have a power and appropriate characteristics that do exactly what you want. Thus, I don't see a need for another one.  It probably costs more than you want it to cost, but it's so stupidly powerful that it SHOULD cost a ton, IMHO.  Even your two-stage example of inability to lose limbs is actually covered by Regeneration; all you need to do is use the optional Hit Location rules to determine which limbs were hit, and consider them disabled or destroyed if enough BODY is done … then let Regeneration with Limb Healing ability do its thing to bring them back.  No special tables or limitations needed, at all … and no need to create a custom power around it; just use the optional rules you already have around Hit Locations in conjunction with Regen and you're golden.
 
That said, if you insist that this isn't a special effect of Regeneration and are absolutely intent on building some janky house-rule power around it, then I believe that the cost to achieve what you want using Regeneration and Recovery should be your guide as to the appropriate cost of whatever you build.  i.e. I think this should cost far more than the Automaton powers tend to cost if it's to be near-instantaneous.  To put that into perspective, Recovering 20 BODY per turn with the ability to Heal Limbs and Resurrect costs … (wait for it) … 345 active points.  So that's the kind of power and pricing you're talking about when you want someone to be unkillable and be able to get up and walk away from Wile E. Coyote-like deaths in Hero in 12 seconds.  (And that doesn't include the cost of a high REC to go with it...)

Appropriate amounts of per-turn Regeneration (with the ability to resurrect and heal limbs) fail to properly represent this how, exactly?  I ask because the ability to be Toon-like 'bonked' is effectively a special effect of the ability to die and recover from it.  That's Regeneration in Hero … and a LOT of it at a very high cost if it's supposed to be near-instantaneous.  Do note that even Wile E. Coyote peels himself off of rocks or walks away looking like an accordian after being pancaked … taking roughly a Turn (12s) to do it as he exits a scene in an amusing way … so I don't think you need to recover faster than that.
 
You probably need a high REC to go with it so that you have some END and STUN, of course … unless you forsee the character needing to take it easy like Wile E. Coyote does (when he goes back to the drawing board) before it goes back into action.
 
Put another way, I think we already have a power and appropriate characteristics that do exactly what you want. Thus, I don't see a need for another one.  It probably costs more than you want it to cost, but it's so stupidly powerful that it SHOULD cost a ton, IMHO.  Even your two-stage example of inability to lose limbs is actually covered by Regeneration; all you need to do is use the optional Hit Location rules to determine which limbs were hit, and consider them disabled or destroyed if enough BODY is done … then let Regeneration with Limb Healing ability do its thing to bring them back.  No special tables or limitations needed, at all … and no need to create a custom power around it; just use the optional rules you already have around Hit Locations in conjunction with Regen and you're golden.
 
That said, if you insist that this isn't a special effect of Regeneration and are absolutely intent on building some janky house-rule power around it, then I believe that the cost to achieve what you want using Regeneration and Recovery should be your guide as to the appropriate cost of whatever you build.  i.e. I think this should cost far more than the Automaton powers tend to cost if it's to be near-instantaneous.  To put that into perspective, Recovering 20 BODY per turn with the ability to Heal Limbs and Resurrect costs … (wait for it) … 345 active points.  So that's the kind of power and pricing you're talking about when you want someone to be unkillable and be able to get up and walk away from Wile E. Coyote-like deaths in Hero in 12 seconds.  (And that doesn't include the cost of a high REC to go with it...)