keybase / keybase/keybase-issues
Request: Ability to delete exploding messages
- Dominant language
- No language data
- Stars
- 899
- Forks
- 40
- PR merge metrics
- No merged PRs in 30d
Description
It was good to see the introduction of exploding messages, and I'm sure this took quite some work to implement. However, at present users are not able to *delete* exploding messages they have sent (they can only explode them).
This is an issue because while the clear and cipher-text is thrown away, a record of the message is kept which at present can only be erased by scrubbing the entire chat. By contrast, regular messages can be deleted by their author at any time without leaving a trace. This persistent record of an exploding message seems very undesirable in the event that an attacker gained access to the chat, since now they would know not only what was in the conversation, but also that other things had also been discussed. Moreover, given that given you envisage that exploding messages be used for [your most *intimate* and most *corporate* messages](https://keybase.io/blog/keybase-exploding-messages), it would be logical for the attacker to infer that there was even more sensitive information in those messages.
It may be that this is an unlikely scenario, and I appreciate that the current design helps maintain a conversational flow which include transient messages, but it would be preferable that:
a. Users were able to delete their own exploding messages, both pre- and post-explosion; and
b. By design, record of exploding messages was erased, perhaps after a user-configurable time period.
Contributor guide
No contributing guide indexed for this repository
Research direction
No files, tests, or entry points are identified in the issue. First review the existing exploding-message implementation and deletion behavior, then clarify whether both requested deletion cases and configurable record retention are in scope. Done should include a decided design and tests covering pre- and post-explosion deletion.
Written by the indexing model from the issue text.
Assessment
- Domain
- security
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100