Here are a few idea I have. From past experience I’m sure they match the average GSoC participant ability to deliver. I also know my ideas are not sexy at all and do not attract many students
Rewrite ansible based package creation for the SecureDrop app into proper Debian GNU/Linux packages and submit them to Debian GNU/Linux and have them in tails by default
Replace codecov with a coverage based Free Software alternative by upstreaming the corresponding functionalities to coverage
Improve monitoring of SecureDrop source interfaces (Python + Django + stem) - We currently use nagios to monitor source interfaces. We send alerts to SecureDrop administrators when we detect their source interfaces are down. Unfortunately the current approach leads to a lot of false positives, which causes admin frustration and confusion. The student’s project would be to develop an improved monitoring solution using the Tor stem library and integrate it into the securedrop.org directory (being migrated to Django).
Just in terms of general strategy in picking projects that are appropriate for students: I think it’s best to have students work on useful work (including prototyping work) that the core team does not have the bandwidth to take on, but not have students work on critical work on the roadmap as it creates unnecessary pressure (e.g. moving to Xenial).