Review Board's focus historically has been to provide tools for

automatically or manually catching problems with code before it goes

into a product. The automatic checks come by way of extensions like

Review Bot, and these can only check code once it's been published.
There are issues that are best caught before code goes up for review,

and there are also issues that are best caught by default without

needing to install a tool like Review Bot.
Two examples would be credentials accidentally left in code and

so-called "Trojan Source" attacks (where code is either accidentally or

intentionally added to a file that displays one way to a user but

executes another way).
This begins laying the foundation for code safety checkers, which can

look for suspicious content in code before it's ready for review,

flagging issues that are found.
This will be used in the diff validation API and the diff viewer to

highlight any issues that are found. Specific code safety checkers will

be implemented in future changes.

Unit tests pass.
Tested this along with a code checker implementation and the upcoming

diff viewer updates.