Thanks for the hint! Can’t remember to see that before at the Github commits page.I’ll keep my eyes more open in future 
Definition of „Critical“:
I‘d like to cover situations of potential risks (= probability x amount of danger) for user data and system / security integrity. The threshold should be defined by security experts in your team, with a good understanding of FMEA at Discourse.
More general: It’s nessary to categorize “unwanted effects” in terms of quality management.
Unfortunately, quality management never was and will be a popular topic due to the extra classification and documentation work. But I think, any QM is better than none.
Or did I miss something at the agile development culture of Discourse? 
