For a while, Metasploit has not used the old database-backed module cache in favor of the lightweight JSON data store. This also means that the db_rebuild_cache command has been broken.
While the base module cache usually stays up to date, if you delete a module as a developer, there's currently no great way to make the search function forget about that module unless you rebuild the cache manually (a procedure mostly documented inside of an automated build job).
This moves the logic from that build job into the a new reload_search command, and deprecates the old one.
And take the Jaden Smith approach, as @busterb quipped to me. :)
This one's a little weird, since you normally import scans into
Metasploit, but now that creds are first-class in the database, it makes
more sense to be able to import them.
Currently, your alternatives are post/windows/gather/credentials/gpp,
which requires a session, and auxiliary/scanner/smb/smb_enum_gpp, which
requires a network scan.
The -S flag for console commands, backed by search functionality
in Rex' tables, originally pushed upstream in #1604 (iirc), lacks
coverage for a number of commands which benefit a good deal from
inline filtering of the potentially large number of results.
Push more -S flags and surrounding table functionality upstream
to provide coverage for the console commands included in framework.
Include a fix for deleting hosts when DB references are a problem.
Include a fix for the upstream route command wherein scope must be
defined for the routing target by assuming a /32 without explicit
definition.
Note:
With this in place, console behavior when filtering results is
roughly analagous to the R7 filtering in web UI, which should help
those of us trying to use both maintain corresponding workflows.
Testing:
Used in-house for years, though changes to the diff from upstream
and our fork (expunging some internal code) are untested, so would
appreciate eyes and hands on.
Brent Cook