RE: CVE-2024-52005 status?
- From
rsbecker@nexbridge.com <rsbecker@nexbridge.com>
- Date
- Nov 2, 2025, 16:19 UTC
- Message-ID
- <01a101dc4c14$7f8f8000$7eae8000$@nexbridge.com>
- In-Reply-To
- <aQd_iisOrwX909Fr@fruit.crustytoothpaste.net>
On November 2, 2025 10:58 AM, brian m. carlson wrote:
Show 30 quoted lines
>On 2025-11-02 at 00:48:53, Thomas Klausner wrote: >> Hi! >> >> I'm trying to find out more details about CVE-2024-52005 so I can >> handle it correctly in pkgsrc. >> >> https://nvd.nist.gov/vuln/detail/CVE-2024-52005 >> >> points to >> >> https://github.com/git/git/security/advisories/GHSA-7jjc-gg6m-3329 >> >> which lists many affected versions, but no fixed versions. I only >> found one reference to the CVE in the release announcement for 'Git >> for Windows' 2.47.1(2): >> >> https://groups.google.com/g/git-packagers/c/ZWDc6K6tYcc/m/VxikezMeCgAJ >> >> The CVE description sounds like this also might affect non-Windows Git >> versions. Can someone please tell me if this was fixed, and in which >> version? > >We determined this was not a vulnerability. The Git for Windows maintainer >disagreed and listed it as one nevertheless. > >The problem is that on some terminal emulators, including at least some of the ones >available on Windows, you can execute arbitrary binaries by creating a hyperlink via >terminal codes and having someone click on it. >That is, however, a bug in the terminal emulator and not in Git, since clicking a URL >should not lead to code execution.
If the code that dumps messages like this is localized, perhaps we can run a filter that removes any non-printable characters (like \ESC) that could trigger this. It might be switchable in some variable like core.filterUnprintableCharacters=true to opt-in to prevent this. However, some emulators might just see a URL and treat it like a link, so I don't know.