Skip to content

Product Meeting: Sanitization and Export

Allie Crevier edited this page Jun 10, 2021 · 6 revisions

Here are some brainstorming notes from our product meeting on June 10, 2021. See https://docs.google.com/document/d/1UNDnGpbfrm-iLAH85hEGIfo87bd6AI6ny8lQd8za4Eo/ for more ideas and questions around this topic.


How do we see post-triage (after users have decided whether a submission is actionable) workflows working for more than just our current pilot groups?

  • sanitization is a must, because the very next step is going to be "let's share this with editor/teammates/research". might even need to share with general counsel, and in which case, there will probably be less of an apetite to transport derivatives safely from the workstation.
  • One distinction between orgs is just the sheer submission volume -- "triage" is not really a problem for orgs that only get one submission a month
  • Technical skills and risk-awareness vary greatly! Many journalists will not be aware at all of malware/metadata risks when managing documents+1 to this, should aim to make sanitization an easy default +1 (sometimes a journalist might want to choose which app to use to sanitize/mitigate malware and sometimes a journalist might want this to happen automatically in qubes: an automatic sanitization pipeline every time you export to usb or vm)
    • barebones implementat first with desire paths
    • Metadata might be part of a tip?
    • important to keep derivative copies and making it apparent where the original is
  • some orgs will still be physical rather than remote, so support for print workflows still important (though with WS folks may adapt working habits) +1
    • do we have enough hardware support for this?
  • Following sanitization, we ought to think about the safety of the channels where people are exporting. For example, if someone exports to Signal, then sends off to a compromised device… Well, that's no good.
  • Questions about how to recreate/redact documents (e.g., clasified gov. docs or internal corporate docs that may implicate a source if published in their original form/entirety), OCR PDFs, etc. within Qubes once a spicy doc is found to be actionable
  • I think most are going to mostly focus on getting the good leaks out of an unfamiliar Qubes environment into their own Mac/Windows computer quickly. Hopefully safely too but they won't necessarily know what sanitization means 'til training shows them. +1, from prior user interview where user went from SD to mac "to sanitize on mac"

What would be good resources/tools for post-processing (sanitization, redaction, detection)?