Re: [gentoo-user] app-misc/ca-certificates

[William Kenworthy]

On 1/6/21 9:29 pm, Rich Freeman wrote:
> On Tue, Jun 1, 2021 at 7:59 AM Adam Carter <adamcarter3@×××××.com> wrote:
>>> And another "wondering" - all the warnings about trusting self signed
>>> certs seem a bit self serving. Yes, they are trying to certify who you
>>> are, but at the expense of probably allowing access to your
>>> communications by "authorised parties" (such as commercial entities
>>> purchasing access for MITM access - e.g. certain router/firewall
>>> companies doing deep inspection of SSL via resigning or owning both end
>>> points).
>> AFAIK in an enterprise MITM works by having a local CA added to the cert stores of the workstation fleet, and having that CA auto generate the certs for MITM. That didn't work with certificate pinning, but pinning has been deprecated.
> So, I don't know all the ways that pinning is implemented, but if
> you're talking about using MITM to snoop on enterprise devices on the
> enterprise network I'd think that pinning wouldn't be an issue,
> because you control the devices from cradle to grave.  Just ensure the
> pinned certificates are the ones that let you MITM the connections.
>
> Now, if your organization has some sort of guest network for
> non-enterprise devices then pinning would obviously block MITM of
> connections made by those devices.  Really though I'm not sure you'd
> want to be snooping stuff like this - it seems like more legal
> headaches than it is worth.  You want to sniff your OWN traffic for
> IDS/etc or other unauthorized use, and since you're sniffing traffic
> from devices you own you don't have the same legal issues (I won't say
> no legal issues, but certainly monitoring your own devices is very
> different from monitoring those you don't own).  You shouldn't even be
> allowing uncontrolled devices on those networks in the first place.
> If you want to detect unauthorized devices MITM isn't really the best
> solution - just use positive authentication of known-good devices
> up-front and anything that doesn't pass that test is treated as a
> threat and shouldn't even be able to send traffic.

When discussing what traffic is looked at in an educational setting it
looked like the system examined everything except mainline banking URL's

For OpenVPN through a MiTM SSL proxy: Double wrap in SSL - outer one
uses their cert so it does not fail that test - inner one uses your self
signed cert for OpenVPN running on port 443 TCP.  At the destination use
the sslh multiplexor to divert SSL to stunnel/second sslh instance etc.
to strip the SSL wrapping appropriately. Works using a combination of
proxytunnel on the Windows side and stunnel on the linux end if needed -
very flexible).  There are are a few other enhancements for pinholing
more  difficult sites.  Performance is entirely adequate for a road
warrior setup when travelling (via a Raspberry Pi AP).  I have had to
get a lot more sophisticated than back in the day when httptunnel was
all that was needed :)

BillK