[94] reintroduce scheduled safety number re-trusting job

Closes #94 (closed)

Context

• #70 (closed) introduced a scheduled job that would retrust every signalboost user's number every 24 hours as a stop-gap for figuring out more fine-grained fixes to key rollover problems
• #88 (closed) rolled that change back since it caused signald to fallover
• this MR reintroduces the job to test the hypothesis that building from source at current upstream repo head (which we do as of #83 (closed)) will resolve the issue because it includes an upstream commit that eliminated a null-pointer-error that likely was causing the crash. (i.e., this commit: https://gitlab.com/thefinn93/signald/commit/dbf32b6c3a33608268f606a5faf6a4e293c36f60)