sssoleileraaa
commented
2 years ago @eloquence, I agree that the threat model impact is minimal. We'll also have a security review after some ideas have been prototyped and after we've written an engineering design doc. Now that the bulk of Qubes 4.1 migration work is complete, I think we can pull this issue into the near-term column on our project board (https://github.com/orgs/freedomofpress/projects/1#column-2333132). We'll want to start thinking about this early from a product perspective between @tina-ux, @gonzalo-bulnes, and myself. We'll also want to consider team capacity with the limited number of developers that would be available to work on this (we've also been talking about prioritizing https://github.com/freedomofpress/securedrop-client/issues/1349 (Export to VM/ Sanitization).
Any thoughts before I throw this onto the project board (which I believe we're planning to review and prioritize mid August)?
Description
A feature to define a list of patterns or terms that, when found in a source message, cause the source to be flagged as spam (potentially rendered in a separate list).
How will this impact SecureDrop users?
All evidence from user research and support indicates that spam is an increasing problem for journalists and administrators.
How would this affect the SecureDrop Workstation threat model?
Because the SecureDrop Client already stores all submissions and messages in clear text, the threat model impact seems limited. We would need to decide if the keyword list can be synchronized with the server.
User Stories
As a journalist, I want to be able to quickly distinguish spam submissions from valid ones, so I can focus most of my attention on useful submissions.