Do persistent area-effect damage spells break Compelled Duel?

Question

Situation: Paula the Paladin and her friend Severus the Sorcerer are fighting a monster. Severus casts cloud of daggers (PHB, p. 222) in front of (but not in contact with) the monster:

You fill the air with spinning daggers in a cube 5 feet on each side, centered on a point you choose within range. A creature takes 4d4 slashing damage when it enters the spell's area for the first time on a turn or starts its turn there.

Then it's Paula's turn. She stands behind the cloud of daggers and casts compelled duel, which says in part (PHB, p. 224):

The spell ends if you attack any other creature, if you cast a spell that targets a hostile creature other than the target, if a creature friendly to you damages the target or casts a harmful spell on it, or if you end your turn more than 30 feet away from the target.

The monster then runs through the cloud of daggers and takes damage.¹ Does compelled duel end?

The "casts a harmful spell" condition is almost surely not relevant; Severus cast the spell before Paula cast compelled duel, and he didn't cast it on the monster, but on an area adjacent to the monster. I'm much less sure about the "damages the target" condition.

On the one hand, he cast a damaging spell which then damaged the monster. Usually we'd consider that doing damage to it, even if the monster is only hit by being in an area of effect.

On the other hand, Severus didn't do anything. He had previously created a cloud of daggers somewhere in the room. The immediate cause of the damage is the daggers; the proximate cause is the monster's decision to run through the cloud.

(This question focuses on cloud of daggers because that's what happened at the table. I assume the logic would be the same for any spell that has a non-instant duration and creates a damaging area of effect (cloudkill, moonbeam, wall of fire, etc.); if not, an explanation of why not would be welcome.)

---

¹ Severus is aware that compelled duel doesn't force the monster to mindlessly charge at its designated target, but he's betting on the monster being stupid enough to do that; possibly there isn't any other route to get to Paula.

Answer 1

The spell ends

This is still kind of tricky, but I think the language of compelled duel (PHB, 224) gives us enough information regarding what ends the spell to include this scenario:

Let's go through the list (as you did!). The full set of ending conditions is:

The spell ends if you attack any other creature, if you cast a spell that targets a hostile creature other than the target, if a creature friendly to you damages the target or casts a harmful spell on it, or if you end your turn more than 30 feet away from the target.

1. The spell ends if you attack any other creature... No, you did not attack another creature.
2. ...if you cast a spell that targets a hostile creature other than the target... No, you did not do this.
3. ...if a creature friendly to you damages the target... Yes! A creature friendly to you damages the target. Their spell damaged the target, gate has been passed and spell ends.
4. ...or casts a harmful spell on it... No, they did not cast it on the creature.
5. ...if you end your turn more than 30' away from the target. No, you did not do this.

The target was damaged by an ally

Gate 3 is the problematic gate as the spell your ally cast damaged the target, and therefore your ally damaged the target. Spells and their effects are not separated from the caster. This is especially true for anything requiring concentration which provides a very real and direct link between the ongoing spell and the caster.

Just to clarify, if there was a natural and pre-existing terrain that caused damage as the target made it's way to you, that would NOT end the spell.

Answer 2

RAW, Most Likely No

The question isn't specifically addressed here, but a question on the Sage Advice rules answers column strongly implies it does not. Here's the relevant text, answered in regards to moonbeam but specifically mentioning cloud of daggers as being similar:

[...] Our design intent for such spells is this: a creature enters the area of effect when the creature passes into it. Creating the area of effect on the creature or moving it onto the creature doesn’t count. If the creature is still in the area at the start of its turn, it is subjected to the area’s effect.

Entering such an area of effect needn’t be voluntary, unless a spell says otherwise. You can, therefore, hurl a creature into the area with a spell like thunderwave. We consider that clever play, not an imbalance, so hurl away! Keep in mind, however, that a creature is subjected to such an area of effect only the first time it enters the area on a turn. You can’t move a creature in and out of it to damage it over and over again on the same turn.

In summary, a spell like moonbeam affects a creature when the creature passes into the spell’s area of effect and when the creature starts its turn there. You’re essentially creating a hazard on the battlefield.

(emphasis mine)

If the designers' intent was to have such spells create a hazard on the battlefield, it stands to reason that the effects of the spell are not damage imposed by a creature friendly to you, but rather by a hazard on the battlefield.

An important distinction is that any damage that occurs happens at the start of the target's turn or on the turn it enters the spell effect, and not during the caster's turn. Another strong bit of evidence can be found on page 194 of the Player's Handbook, under "Making an Attack":

If there’s ever any question whether something you’re doing counts as an attack, the rule is simple: if you’re making an attack roll, you’re making an attack.

Answer 3

Yes, the spell should end

That the damage from cloud of daggers (and similar effects) is caused by the caster is relevant for that damage benefiting from or triggering several class features, like the Phoenix Sorcerer's Mantle of Flame or the Necromancer Wizard's Grim Harvest. Generally, these features are weak and need all the help they can get, and damage over time spells like cloud of daggers are uniquely poised to benefit from said features. By ruling that the damage done this way isn't dealt by the caster, some already powerful options are made better while these weaker class features are made worse. Consequently, given that the rules do not clearly indicate either way, I believe that it would be better to rule that this damage is, in fact, dealt by the caster.