Do you use skill challenges?

I would not stick to the old Skill Challenge pattern, but I might use something similar. Consider the old "Arcane Lock" situation. That is:

Arcane Lock said:

There's a very complex magical lock on a vault door. There are, in fact, five separate locks preventing the lock from being opened. A warning in Draconic reveals that all five locks must be opened in sequence or a trap will be triggered. To open the vault, the PCs must succeed on five Intelligence (Arcana) checks before 2 failures. If the PCs fail, the locks reset and 2d4 goblins arrive.

Click to expand...

I would probably not bother with this design now. It's not a particularly interesting challenge to make the players roll over and over for the same locks. It's easier to just have them roll once. I feel like all we're adding here is more dice rolling to smooth out variance. That's... not interesting. It's really better than:

Arcane Lock no Skill Challenge said:

There's a very complex magical lock on a vault door. There are, in fact, five separate locks preventing the lock from being opened. Unlocking a lock requires a successful Intelligence (Arcana) check. Up to one player may attempt to unlock a lock in each round and there five locks in total. If a PC fails to open a lock by 5 or more, an alarm is triggered. On the following round and every 1d4 rounds thereafter, 2d4 goblins swarm into the room and attack any creatures they can see that are foreign to them. The goblins continue to arrive even if the alarm is somehow disabled. A total of 8d4 (20) goblins can respond to the alarm.

Click to expand...

Or the "Lost Navigation" situation:

Lost Navigation said:

If the players attempt to navigate the swamp, they will find it is deceptively treacherous. The PCs must succeed on five Wisdom (Survival) checks before 2 failures to navigate the swamp. If the PCs fail, roll once on the random encounter table. The PCs are forced to retreat and begin again.

Click to expand...

In 4e, failure consumed a healing surge instead of triggering a roll on the encounter table, so failure was even less eventful for the players than a random encounter.

I'd rather do this:

Lost Navigation no Skill Challenge said:

If the players attempt to navigate the swamp, they will find it is deceptively treacherous. The swamp is 15 to 20 miles across and is mostly difficult terrain. Every hour of travel (~1.5 miles without a Ranger guide, ~3 miles with a Ranger guide), the PCs must succeed on a Wisdom (Survival) check to determine if they have successfully navigated that leg of the swamp. If they fail, the PCs lose a mile of progress and they must double back and go around an impassable or dangerous area of the swamp (so 0.5 miles or 2 miles of progress). Roll once on the random encounter table as the PCs fall prey to determine the nature of the hazards the party has encountered. A result of "no encounter" means the party has encountered a section of the swamp that is impassible to most characters (sinking sands, impenetrable briars, open water covering thick and deep mud, etc.).

Click to expand...

Sure, it's not open ended for endless failure, but I don't want my PCs stuck in a swamp all night. And instead of forcing the challenge into a fixed pattern, it can just be designed to feel correct for the situation. I mean, yes, that's realistic that they might get lost for days in a swamp, but it really isn't fun or interesting or what my players want to experience.

Finally, I absolutely want to avoid the dreaded social Skill Challenge:

Convince the Mayor said:

The mayor can be convinced to aid the party, but he is stubborn. The PCs must succeed on five Charisma (Diplomacy) checks before 3 failures. If the PCs succeed, the mayor provides the party with The Map and instructions on where to find the MacGuffin.

Click to expand...

I saw so many skill challenges like this in 4e. It commits the sins of 1) Requiring the PCs to succeed to progress with the adventure (i.e., if you fail, you can try again indefinitely), 2) Discouraging role-play in favor of dice rolling to solve every challenge, (I don't need less reasons to role play and more reasons to roll dice! We have combat for that!) 3) Ignoring the possibility of arguments so convincing that the encounter should end. ("The goblins that have this MacGuffin also took your only daughter." "Not good enough, I need 4 more reasons.")

I still remember one 4e Skill Challenge in a module from Dungeon where the party is expected to convince some mayor or prince to help them. If the party fails, the module provides no guidance. The only way forward is to convince the mayor. If the party suceeds, they're still told to go find evidence and come back. When the PCs do that, the the party has to complete another skill challenge to convince him even with the evidence. Worse, the DCs for convincing the mayor/prince were higher after you had evidence because of the way DCs scaled with level in 4e. And, again, if the party failed, their only recourse was to try again. They had to have the mayor or prince's help to reach the next leg of the adventure. It was just everything I hated about Skill Challenges wrapped up in a single package. It was all about checking the mechanical boxes of having enough skill challenges to satisfy WotC so they could show off their brand new mechanic, and nothing about building a compelling or sensible narrative. That has left such a horrible taste in my mouth that I simply don't like the skill challenge mechanic.

Morrus said:

2. Each PC states what they are doing towards the skill challenge. If it makes sense to the DM, they make a check.

Click to expand...

This is the reason I stopped using them. Sometimes a PC just doesn't have any way to meaningfully impact a challenge, other than to fail it faster. Roping them into situation like that is just unfair for the player individually and the group as a whole.

I do not use skill challenges. As a player, they always (always) feel contrived, and as a dungeon master, 5th edition's action-focused narrative emphasis brings the story to life in ways that dice-focused game emphasis serves only to undermine.

One of the Ashes of Athas modules had a Skill Challenge that my character happened to be optimized for. Other players (who were much better in combat than I) had to roll a 20 to avoid failure ! They weren't interested, because it looked to them like "you are doomed auto-fail and you will drag the group down with you."

When I write my own material, I try to provide something like a Skill Challenge, although I don't use the mechanic or description; I just present the PCs with a situation where "I hit it with my sword" is clearly inappropriate. Some players try anyways -sigh- .
Usually I will indicate in my notes "hard", "medium", and "easy" levels for different skills. Sir UselessOrnamentUponSociety responds well to Diplomacy (medium) but gets his hackles up when somebody tries to Intimidate him (hard). But to really get his attention, bring him by a window and have your Bard play his family's ancestral anthem (if you have a copy of the sheet music, easy Perform) out on the lawn. Of course some NPC will 'clue in' the PCs ... but they don't HAVE to take the easy route ...

I found out that newer players can get into a skill challenge more when you encourage them to use their imaginations, and/or tell them to think of how they would solve a real-world analogy (broken wagon = flashing idiot light on car dashboard).

Morrus said:

Yeah; mine is definitely a player-narrative slanted approach. I like to do that from time to time, and it's so easy to do impromptu.

Click to expand...

What do you do when a player makes a sucessfull skill check but then describes a course of action that will not help in the least for the overall challenge?

In 4e, sure, though I usually expand on them.
Before that, obviously not - I might have had non-combat challenges, but they'd've been completed conventionally for the ed, be that casting the right spell, saying the right thing, solving a puzzle, or using a special ability or skill. Mostly pretty arbitrary.
5e could be used with an SC, it retained 4e's relatively small swing between trained and untrained, but the open structure doesn't play well with 5e, which otherwise works best keeping as much as possible behind the screen. Puzzle style challenges fit the 5e classic vibe much better, they could include checks, but the main point would be figuring out what to do, not how well the PC carries it off....

After previous experiences with Skill Challenges I will never use them again.

Yep, and they're more or less along the same lines:
1: Determine what can and cannot be done to resolve a situation. IE: diplomacy isn't going to move rocks.
2: Determine which methods of resolution will be easy/moderate/hard. EX: Diplomacy may convince the neighboring Ogre to move rocks, but it won't be easy.
3: Reward players who take a multi-pronged approach to a situation. Which means: Don't forget to let the villagers know you're bringing an Ogre into town to help them fix the dam.
4: Be prepared for your players to think up some resolution you didn't expect.
5: Reward players for overcoming the situation: it doesn't matter if they did it the way you want, the easy way, the long way, the hard way. Same XP reward.

I've read about skill challenges but never tried running any. I find the general idea interesting and a good way to showcase characters and their areas of expertise but I'm not sure I'd be able to run them so that they feel like a smooth part of gameplay.

Any time you make death saving throws you're basically playing a sort of skill challenge, without the skill part.