This is a pretty sensitive new script. As said during our sprint in January, I would like at least 1 security expert to look into it.
I remember someone (segfault?) asked jvoisin to do so on XMPP, but I don't know if there's been any follow-up to this.
So I'm assigning to FT for now, to handle the communication part, which I suppose may look like this:
briefly document the intended use and threat model here
have jvoisin audit
follow-up on jvoisin's audit (!666 (merged))
answer #18373 (comment 176542)
Blocked by #18360 (closed)