Question from an online list: What are some collaborative student-student activities that lend themselves to a language acquisition method based on dense, high-quality comprehensible input?

You got me.
Student to student input is not comprehensible input, it’s “interput”. That’s low-quality forced interaction between people who don’t know the language yet, like students.
Students CAN provide input — but only of the portion of the target language they have really, really, REALLY acquired. Usually this doesn’t work out very well, for the same reasons that people’s demonstrations of TPRS in languages they really don’t speak to at least a reasonable conversational level are wooden and less than effective, and often produces students who think they “know” that language, but are making enormous basic errors. Neither is able to give flexible comprehensible input in the target language, and neither knows if he is saying something incorrectly, so there might — might — be some benefit if the stars line up right, but it’s really anyone’s guess whether that happens or not.
Now, student to student interaction can be okay for accomplishing tasks or doing activities in the TL — ┬ábut there will not be any new acquisition of correct language that the students have not yet acquired, because where is it going to come from, if the students don’t have it already? Reading it off a piece of paper is not the highest quality input available, unless you get quality points for checking off Danielson rubric boxes.
There are many legitimate reasons to do interput activities or straight output activities. NONE of them are “to help students acquire new language in the target language”.