There are two separate things wrong with Halberd. One is the claw. The other is that it shortens games.
Anyone who starts talking about how I should chill out about laser or cannon makes me feel like they have less stage knowledge than I do about Halberd or believe the reverse of that is true, perhaps a presumption based on my being a stage conservative or having a "wrong" opinion. In advanced play on Halberd with stage knowledge on both sides, the claw has like ten times the impact of laser and cannon combined due to the greatest amount of opportunity for an opponent's exploitation. It is not worthy of any time of discussion compared to the claw.
The claw randomly selects one player to be struck by it. Neither player has any control over that selection. The player selected by the claw always has a strategic disadvantage, at a minimum. You have to use invulnerability or blocking to avoid taking damage from the claw, and your opponent knows that you have to use invulnerability and blocking to avoid taking damage. Good players can capitalize on that and punish the predictable behavior. It's also possible to start a juggle before the claw selects you and be forced to stop the juggle to avoid the claw.
Arguing that the claw doesn't always kill is a terrible argument, the only acceptable argument would be that somehow good players can totally avoid the impact of the claw consistently, which would also be bunk, but less bunk. Once the claw has dealt its chunk of damage and put one player into disadvantaged state, it's already had a huge impact on the game from RNG.
The second is the shorter games. Halberd is the stage that makes KOs earlier more so than any other stage. This reduces the total amount of interaction between the two players, sometimes by an entire's stock's worth. This happens to amplify the impact of the claw, although either issue would be a problem on its own.
If Halberd were legal, I would want it to be a 4 stock, 8 minutes stage. That would be if I could be talked into tolerating the claw. Players should get a consistent amount of skill testing in their sets, whether or not correct strategy led them to striking down to Halberd. Players who know they are stronger than other players should not feel obligated to strike a stage that will make it easier for the weaker to win just by virtue of behaving like a 2 stock match, either. That reduces the sense of efficacy players get from practicing and mastering smash beyond other players by giving some underdog edge from the free strike.
Stages that last longer than other stages are ok, they actually test skill more than normal. Testing skill more than normal is obviously ok, it's not symmetrical to testing skill less than normal at all. Time out games can be incredibly skill testing. We set the stock count to 3 to make sure that game is at least long enough for good skill testing, then set the timer to 8 so that the game is at least short enough for the tournament to finish on time. A stage that promotes slow play is unable to modify the 8 minute timer (if a transformation stage lagged the timer or something, then that would be what it takes to make the "if short is bad then long is bad too" symmetry that AA referred to). A stage that promotes fast play can rather easily remove a stock or part of a stock's worth of interaction from a game.
As for the long list of stages you referred to, AA, yes, there's some stages that seem like they are banned by the hivemind for a style of play we don't like. But in actuality, a lot of people don't enjoy them because their games feel shorter, and people don't like when a 3 stock game feels 2 stock, but they can't always put their finger on that. Walkoff stages often seem that way, where lots of early kills near the edges makes the game feel shorter, then people think the stage is annoying and unfun and they can't explain exactly why or they pin it on something else, but really they're just upset that the stage is genuinely less competitive and rewarding the more skillful player less. I think walkoff stages really do degenerate into walk-off camping eventually with or without Brawl D3, but even without the tactic in its purest form, you can play somewhat closer to the walkoff edges against a more skillful player and more easily get an upset. Super Mario Bros. scrollers are similar. I think the thing with Skyworld is onstage spikes shooting through the clouds as a special mechanic, which would be another game shortener.
Luigi's Mansion and Wiley's Castle are ones in your list that don't shorten games as an auxillary effect to whatever it is that seems to make people hate that style of stage. With those, I would indeed ride along with your argument ad-absurdum and speculate that they should perhaps actually be legal. I'm not saying that just to keep up with you, I think I'm timestamped for speculating the Wiley should be legal before you made that post. Lmansion cave of life is a game extender, which as I will stick to, is more competitive, not less competitive. I think those stages actually might have more merit than some of the ones currently accepted as legal that have shorter gameplay. These are both highly polar stages which makes it very important that you are using a system with plenty of bans for game 2 and 3, but they do not trivialize skill testing like Halberd.
"Also it has one of the lowest ceilings in the game but that's not a point against it, that's a valid reason to strike or CP it depending on the character and matchup." - ParanoidDrone
The issue is not Kirby vs. Pikachu, where Kirby knows he kills off the side, Pikachu knows he kills off the top, and Kirby has to strike Halberd. That's fine. That's strategy. That's good. If I wanted to forbid that kind of interaction I'd have to be against Kongo Jungle's high ceiling too to be consistent.
The issue is Pikachu vs. Diddy, where the Pikachu knows he's the best in the state, the Diddy knows he's fifteenth best in the state, they both kill off the top super early, the Pikachu is forced to strike Halberd to make sure he gets a long game (with less RNG) to be certain that his superior skill lands him a win against the weaker player. But Pikachu is now a strike behind, and the Diddy player effectively won a free strike for being bad, and you don't want that.