oss-sec mailing list archives

Re: MITRE is adding data intake to its CVE ID process


From: Peter Bex <peter () more-magic net>
Date: Thu, 9 Feb 2017 15:54:09 +0100

On Thu, Feb 09, 2017 at 02:26:01PM +0000, Jeremy Stanley wrote:
Agreed, having tried to figure out the form it seems geared toward
requesting CVE IDs for vulnerabilities you've found in someone
else's software, and not for maintainers of software to request CVE
IDs for vulnerabilities which have been disclosed to them. The
little detail callout icons for the vendor and product fields link
to the CNA coverage list[0]

I'm also concerned about this.  Last time I asked a MITRE employee
about this as a result of the automated mail that oss-security
sends out when it detects a CVE request, and I did not receive a
reply.

So far I've only requested CVE IDs for projects that have no
obvious CNA.

which in turn instructs, "For open
source software products not listed below, request a CVE ID through
the Distributed Weakness Filing Project[1] CNA." So I guess that's
what our project will be using in the future, or maybe just stop
bothering to obtain CVEs on our own and let the various downstream
redistributors of our software who are themselves CNAs issue them as
needed and then fight over whose is the correct one.
[0] http://cve.mitre.org/cve/request_id.html#cna_coverage
[1] https://docs.google.com/forms/d/e/1FAIpQLSeiY7ldJAx-fjU6eSnXDaX5TB--L1ujCQpmGAKnqBSJOcBShw/viewform

For me, having to use a Google docs form is unacceptable.  I try to
avoid Google in my life as much as possible.  I'd rather avoid
requesting a CVE ID, or let someone else do it, if that's the only
remaining option for non-listed free software projects.

The oss-security list was a reasonably good solution for me.  The only
disadvantage is that time passes between initial announcement and the
CVE ID, because (AFAIK) a CVE request needs to contain a reference to
the announcement as "proof" that the vulnerability is real and not a
duplicate.  In an ideal world, free software project leaders should be
able to request a CVE ID _before_ announcing a vulnerability to their
user base.  If there were some way to register people as project leaders,
the "proof" should not be necessary, they should be able to request a
CVE ID with authority.

This delay between announcement and getting a CVE ID has always bothered
me about CVE requests via oss-security.  On the other hand, I like the
fact that there's a somewhat centralised place to watch for important
and early(!) free software security information.

Cheers,
Peter Bex

Attachment: signature.asc
Description: Digital signature


Current thread: