Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 08 Sep 2014 10:21:10 -0600
From: Kurt Seifried <>
Subject: Re: Python robotframework - tmp vuln

On 08/09/14 09:22 AM, wrote:
>> This is the first of many
> The MITRE CVE team obviously has no objection to your use of the
> oss-security list for raising new discussion topics such as the
> likelihood that a '../tmp/ substring represents a security problem.
> The comments below are only about obtaining CVE assignments from
>> the reason I'm not assigning CVE's for these is this is a side project
> A CVE isn't going to be possible without further analysis explaining
> why a vulnerability exists in the specific case. There can't be an
> expectation that someone at MITRE is already familiar with the
> product, or will read and understand the complete source code as part
> of processing an oss-security message.
> Items that seem to be missing from the original message include:
> 1, Is the "merge('../tmp/passing.xml', '../tmp/failing.xml')"
>    debugging code, or is this code realistically used because a
>    different piece of software has created passing.xml and failing.xml
>    files?

It's part of __main__ so it gets executed.

> 2. If there is a realistic situation in which the
>    "merge('../tmp/passing.xml', '../tmp/failing.xml')" executes, would
>    the cwd realistically be a first-level directory such as the /root
>    or /tmp directory?

yes if you run from within /root or /tmp, the "run from /root" would be
the obvious worry as it would indicate the root user is being used.

> 3. For purposes of risk analysis, is unconstrained use of a ../tmp/
>    pathname always equivalent to unconstrained use of a /tmp/ pathname?

If it ends up using /tmp/ then I would say that's a problem. is it
always a problem? I can't say.

> A possible CVE assignment decision might be:
> A. If a different product came with a test suite containing:
>    test_program > /tmp/merged.xml
> then it could have a CVE because /tmp/merged.xml might be a symlink to
> an important file.
> B. If the test suite were changed to:
>    test_program > ../tmp/merged.xml
> with no constraints, then it could still have a CVE, because some
> people run test suites as root with a cwd of the /root directory.
> C. If ../tmp/ is used in "debugging code" that is intended to be run
> by a developer who understands the appropriate cwd, and this
> "debugging code" is not a "test suite" for users, then there is no CVE
> assignment. Admittedly, there might be cases where the distinction
> between "debugging code" and "test suite" is completely ambiguous.

Kurt Seifried -- Red Hat -- Product Security -- Cloud
PGP A90B F995 7350 148F 66BF 7554 160D 4553 5E26 7993

Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)

Powered by blists - more mailing lists

Your e-mail address:

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Powered by Openwall GNU/*/Linux - Powered by OpenVZ