Fully automatic plugin updates

> I believe it's worth considering the option of reviewing in hindsight rather than screening everything beforehand. We would still check plugins when they are first added, meaning someone malicious would need to develop and upload a legitimate plugin, then later introduce malicious content through an update.
>
> It seems to me that the potential for abuse is pretty low, and if we adopt a three-strikes system,
> Strike 1: Warning
> Strike 2: Plugin removed
> Strike 3: Author blocked
>
> I feel confident that we won't get many, if any, people looking to abuse the system.
>
> This would also free up a lot of maintenance requirements so we could then look into implementing systems that make adding a plugin easier, so we'd be making things easier for the 99% of good actors who just want to share their plugin, while introducing a small risk of malicious content being listed on "official" sources for a short time before it's reported and removed.
>
> I can't imagine anyone would consider that insufficient.

_Originally posted by @Hecter94 in [#1781](https://github.com/endless-sky/endless-sky-plugins/issues/1781#issuecomment-3458844388)_

The issue is a policy proposal and names no files, tests, or entry points. First map the current plugin review and publication process, then define the update, warning, removal, and author-blocking behavior; the work is done only when the automated-update scope and safety rules are specified and implemented.

Written by the indexing model from the issue text.