A flaw was found in ruby-git, where the package is vulnerable to command injection via the git argument. This flaw allows an attacker to set additional flags, which leads to performing command injections.

Affected versions of Git have a vulnerability whereby Git can be tricked into sending private credentials to a host controlled by an attacker. Git uses external "credential helper" programs to store and retrieve passwords or other credentials from secure storage provided by the operating system. Specially-crafted URLs that contain an encoded newline can inject unintended values into the credential helper protocol stream, causing the credential helper to retrieve the password for one server (e.g., good.example.com) for an HTTP request being made to another server (e.g., evil.example.com), resulting in credentials for the former being sent to the latter. There are no restrictions on the relationship between the two, meaning that an attacker can craft a URL that will present stored credentials for any host to a host of their choosing. The vulnerability can be triggered by feeding a malicious URL to git clone. However, the affected URLs look rather suspicious; the likely vector would be through systems which automatically clone URLs not visible to the user, such as Git submodules, or package systems built around Git. The problem has been patched in the versions published on April 14th, 2020, going back to v2.17.x. Anyone wishing to backport the change further can do so by applying commit 9a6bbee (the full release includes extra checks for git fsck, but that commit is sufficient to protect clients against the vulnerability). The patched versions are: 2.17.4, 2.18.3, 2.19.4, 2.20.3, 2.21.2, 2.22.3, 2.23.2, 2.24.2, 2.25.3, 2.26.1.

Git is a revision control system. Prior to versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4, repositories with submodules can be crafted in a way that exploits a bug in Git whereby it can be fooled into writing files not into the submodule's worktree but into a .git/ directory. This allows writing a hook that will be executed while the clone operation is still running, giving the user no opportunity to inspect the code that is being executed. The problem has been patched in versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4. If symbolic link support is disabled in Git (e.g. via git config --global core.symlinks false), the described attack won't work. As always, it is best to avoid cloning repositories from untrusted sources.

Git is a source code management tool. When cloning from a server (or fetching, or pushing), informational or error messages are transported from the remote Git process to the client via the so-called "sideband channel". These messages will be prefixed with "remote:" and printed directly to the standard error output. Typically, this standard error output is connected to a terminal that understands ANSI escape sequences, which Git did not protect against. Most modern terminals support control sequences that can be used by a malicious actor to hide and misrepresent information, or to mislead the user into executing untrusted scripts. As requested on the git-security mailing list, the patches are under discussion on the public mailing list. Users are advised to update as soon as possible. Users unable to upgrade should avoid recursive clones unless they are from trusted sources.

A vulnerability was found in git concerning the git shell. A user who comes over SSH could run an interactive pager by causing it to spawn "git upload-pack --help".

"git-shell" is a restricted login shell that can be used on a server to prevent SSH clients from running any programs except those needed for git fetches and pushes. If you are not running a server, or if your server has not been explicitly configured to use git-shell as a login shell, you are not affected.

Upstream patch:

https://kernel.googlesource.com/pub/scm/git/git/+/3ec804490a265f4c418a321428c12f3f18b7eff5

References:

https://insinuator.net/2017/05/git-shell-bypass-by-abusing-less-cve-2017-8386/ http://public-inbox.org/git/xmqq8tm5ziat.fsf@gitster.mtv.corp.google.com/

Impact

When populating a Git repository's working tree with the contents of Git LFS objects, certain Git LFS commands may write to files visible outside the current Git working tree if symbolic or hard links exist which collide with the paths of files tracked by Git LFS.

Git LFS has resolved this problem by revising the git lfs checkout and git lfs pull commands so that they check for symbolic links in the same manner as performed by Git before writing to files in the working tree. These commands now also remove existing files in the working tree before writing new files in their place.

As well, Git LFS has resolved a problem whereby the git lfs checkout and git lfs pull commands, when run in a bare repository, could write to files visible outside the repository. While a specific and relatively unlikely set of conditions were required for this to occur, it is no longer possible under any circumstances.

Patches

This problem exists in all versions since 0.5.2 and is patched in v3.7.1. All users should upgrade to v3.7.1.

Workarounds

Support for symlinks in Git may be disabled by setting the core.symlinks configuration option to false, after which further clones and fetches will not create symbolic links. However, any symbolic or hard links in existing repositories will still provide the opportunity for Git LFS to write to their targets.

References

- https://github.com/git-lfs/git-lfs/security/advisories/GHSA-6pvw-g552-53c5 - https://nvd.nist.gov/vuln/detail/CVE-2025-26625 - https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2025-26625 - https://github.com/git-lfs/git-lfs/releases/tag/v3.7.1 - git-lfs/git-lfs@5c11ffce9a - git-lfs/git-lfs@0cffe93176 - git-lfs/git-lfs@d02bd13f02

For more information

If there are any questions or comments about this advisory: For general questions, start a discussion in the Git LFS discussion forum. For reports of additional vulnerabilities, please follow the Git LFS security reporting policy.

CVE-2025-48384 is regarding a vulnerability in Git where when reading a config value, Git strips any trailing carriage return and line feed (CRLF). When writing a config entry, values with a trailing CR are not quoted, causing the CR to be lost when the config is later read. When initializing a submodule, if the submodule path contains a trailing CR, the altered path is read resulting in the submodule being checked out to an incorrect location. If a symlink exists that points the altered path to the submodule hooks directory, and the submodule contains an executable post-checkout hook, the script may be unintentionally executed after checkout. GitHub created this CVE on their behalf. The documented Visual Studio updates incorporate updates in Git which address this vulnerability. Please see CVE-2025-48384 for more information.

Buffer overflow in git-checkout-index in GIT before 1.1.5 allows remote attackers to execute arbitrary code via an index file with a long symbolic link.

Stack-based buffer overflow in the (1) diffaddremove and (2) diffchange functions in GIT before 1.5.6.4 might allow local users to execute arbitrary code via a PATH whose length is larger than the system's PATHMAX when running GIT utilities such as git-diff or git-grep.

The web interface in git (gitweb) 1.5.x before 1.5.6 allows remote attackers to execute arbitrary commands via shell metacharacters related to (1) gitsnapshot and (2) gitobject.

Sebastian Krahmer of the SuSE security team discovered a remote command injection flaws in the gitweb, caused by an insufficient checking of the inputs used to build argument to perl's open() function. Remote attacker could use these flaws to run arbitrary commands with the privileges of the web server executing gitweb CGI scripts.

Issues are already fixed upstream in the latest git branches. It seems that the security consequences were not noticed when fixes were applied upstream, as multiple occurrences of the similar flaws were fixed in different upstream versions:

CVE-2008-5517 http://repo.or.cz/w/git.git?a=commitdiff;h=516381d5 (fixes issues in gitsnapshot and gitobject, first occurred in 1.5.6)

CVE-2008-5516 http://repo.or.cz/w/git.git?a=commitdiff;h=c582abae (fixes issue in gitsearch, first occurred in 1.5.5)

All current Fedora packages use version 1.5.6+, so neither of the issue apply to them. EPEL versions should be affected by one or both of the issues.

Git for Windows is the Windows port of Git. Prior to 2.55.0.windows.4, a malicious remote Git server can advertise a bundle URI that reaches transportgetremotebundleuri(), fetchbundleuriinternal(), and copyuritofile() in bundle-uri.c during clone or fetch when transfer.bundleuri=true. Non-HTTP(S) values are treated as local filesystem paths, and file URI prefixes are removed, so a bare UNC path or file URI targeting an attacker-controlled share causes Windows to initiate an outbound SMB connection. This can expose NTLM authentication material to the attacker-selected host. This issue is fixed in version 2.55.0.windows.4.

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS automatically collected Git repository metadata without disabling the repository-local core.fsmonitor setting. If a user opens or uses an attacker-prepared repository whose preserved .git/config sets core.fsmonitor to an attacker-controlled filesystem-monitor helper, Git can execute that helper while Codex collects repository metadata. The helper runs outside Codex's command sandbox and without a user-approval prompt, allowing attacker-controlled code to run with the user's privileges. The code can read, change, or delete the user's files and access other resources available to the user's account. An ordinary Git clone does not preserve the source repository's local .git/config; exploitation requires a repository delivered or copied with that configuration intact.

Git LFS is a Git extension for versioning large files. When Git LFS requests credentials from Git for a remote host, it passes portions of the host's URL to the git-credential(1) command without checking for embedded line-ending control characters, and then sends any credentials it receives back from the Git credential helper to the remote host. By inserting URL-encoded control characters such as line feed (LF) or carriage return (CR) characters into the URL, an attacker may be able to retrieve a user's Git credentials. This problem exists in all previous versions and is patched in v3.6.1. All users should upgrade to v3.6.1. There are no workarounds known at this time.

Git is a fast, scalable, distributed revision control system with an unusually rich command set that provides both high-level operations and full access to internals. When reading a config value, Git strips any trailing carriage return and line feed (CRLF). When writing a config entry, values with a trailing CR are not quoted, causing the CR to be lost when the config is later read. When initializing a submodule, if the submodule path contains a trailing CR, the altered path is read resulting in the submodule being checked out to an incorrect location. If a symlink exists that points the altered path to the submodule hooks directory, and the submodule contains an executable post-checkout hook, the script may be unintentionally executed after checkout. This vulnerability is fixed in v2.43.7, v2.44.4, v2.45.4, v2.46.4, v2.47.3, v2.48.2, v2.49.1, and v2.50.1.

Git is a fast, scalable, distributed revision control system with an unusually rich command set that provides both high-level operations and full access to internals. When cloning a repository Git knows to optionally fetch a bundle advertised by the remote server, which allows the server-side to offload parts of the clone to a CDN. The Git client does not perform sufficient validation of the advertised bundles, which allows the remote side to perform protocol injection. This protocol injection can cause the client to write the fetched bundle to a location controlled by the adversary. The fetched content is fully controlled by the server, which can in the worst case lead to arbitrary code execution. The use of bundle URIs is not enabled by default and can be controlled by the bundle.heuristic config option. Some cases of the vulnerability require that the adversary is in control of where a repository will be cloned to. This either requires social engineering or a recursive clone with submodules. These cases can thus be avoided by disabling recursive clones. This vulnerability is fixed in v2.43.7, v2.44.4, v2.45.4, v2.46.4, v2.47.3, v2.48.2, v2.49.1, and v2.50.1.

Git is distributed revision control system. git log can display commits in an arbitrary format using its --format specifiers. This functionality is also exposed to git archive via the export-subst gitattribute. When processing the padding operators, there is a integer overflow in pretty.c::formatandpadcommit() where a sizet is stored improperly as an int, and then added as an offset to a memcpy(). This overflow can be triggered directly by a user running a command which invokes the commit formatting machinery (e.g., git log --format=...). It may also be triggered indirectly through git archive via the export-subst mechanism, which expands format specifiers inside of files within the repository during a git archive. This integer overflow can result in arbitrary heap writes, which may result in arbitrary code execution. The problem has been patched in the versions published on 2023-01-17, going back to v2.30.7. Users are advised to upgrade. Users who are unable to upgrade should disable git archive in untrusted repositories. If you expose git archive via git daemon, disable it by running git config --global daemon.uploadArch false.

https://github.com/git/git/security/advisories/GHSA-475x-2q3q-hvwq https://github.com/git/git/commit/508386c6c5857b4faa2c3e491f422c98cc69ae76 https://git-scm.com/book/en/v2/Customizing-Git-Git-Attributes#exportsubst https://git-scm.com/docs/pretty-formats#Documentation/pretty-formats.txt-emltltNgttruncltruncmtruncem

Git is distributed revision control system. gitattributes are a mechanism to allow defining attributes for paths. These attributes can be defined by adding a .gitattributes file to the repository, which contains a set of file patterns and the attributes that should be set for paths matching this pattern. When parsing gitattributes, multiple integer overflows can occur when there is a huge number of path patterns, a huge number of attributes for a single pattern, or when the declared attribute names are huge. These overflows can be triggered via a crafted .gitattributes file that may be part of the commit history. Git silently splits lines longer than 2KB when parsing gitattributes from a file, but not when parsing them from the index. Consequentially, the failure mode depends on whether the file exists in the working tree, the index or both. This integer overflow can result in arbitrary heap reads and writes, which may result in remote code execution. The problem has been patched in the versions published on 2023-01-17, going back to v2.30.7. Users are advised to upgrade. There are no known workarounds for this issue.

https://github.com/git/git/security/advisories/GHSA-c738-c5qq-xg89 https://github.com/git/git/commit/508386c6c5857b4faa2c3e491f422c98cc69ae76

Git Credential Manager (GCM) is a secure Git credential helper. Prior to 2.5.0, the Debian package does not set root ownership on installed files. This allows user 1001 on a multi-user system can replace binary and gain other users' privileges. This vulnerability is fixed in 2.5.0.

git-daemon in git 1.4.4.5 through 1.6.3 allows remote attackers to cause a denial of service (infinite loop and CPU consumption) via a request containing extra unrecognized arguments.

gitweb/gitweb.perl in gitweb in Git 1.6.x before 1.6.0.6, 1.5.6.x before 1.5.6.6, 1.5.5.x before 1.5.5.6, 1.5.4.x before 1.5.4.7, and other versions after 1.4.3 allows local repository owners to execute arbitrary commands by modifying the diff.external configuration variable and executing a crafted gitweb query.

Cross-site scripting (XSS) vulnerability in Gitweb 1.7.3.3 and earlier allows remote attackers to inject arbitrary web script or HTML via the (1) f and (2) fp parameters.

Git is a source code management tool. When cloning from a server (or fetching, or pushing), informational or error messages are transported from the remote Git process to the client via the so-called "sideband channel". These messages will be prefixed with "remote:" and printed directly to the standard error output. Typically, this standard error output is connected to a terminal that understands ANSI escape sequences, which Git did not protect against. Most modern terminals support control sequences that can be used by a malicious actor to hide and misrepresent information, or to mislead the user into executing untrusted scripts. As requested on the git-security mailing list, the patches are under discussion on the public mailing list. Users are advised to update as soon as possible. Users unable to upgrade should avoid recursive clones unless they are from trusted sources.

When a user clones an untrusted repository and runs Gitk without additional command arguments, any writable file can be created and truncated. The option "Support per-file encoding" must have been enabled. The operation "Show origin of this line" is affected as well, regardless of the option being enabled or not.

A Git repository can be crafted in such a way that a user who has cloned the repository can be tricked into running any script supplied by the attacker by invoking gitk filename, where filename has a particular structure.

By feeding a crafted input to "git apply", a path outside the working tree can be overwritten as the user who is running "git apply".

Using a specially-crafted repository, Git can be tricked into using its local clone optimization even when using a non-local transport. Though Git will abort local clones whose source $GITDIR/objects directory contains symbolic links (c.f., CVE-2022-39253), the objects directory itself may still be a symbolic link.

These two may be combined to include arbitrary files based on known paths on the victim's filesystem within the malicious repository's working copy, allowing for data exfiltration in a similar manner as CVE-2022-39253.

A security flaw was found in the way git-imap-send command (tool to send a collection of patches from stdin to an IMAP folder) of Git performed IMAP server's SSL x509.v3 certificate validation (server's hostname was previously not verified to match the CN field of the particular certificate). A rogue server could use this flaw to conduct man-in-the-middle (MiTM) attacks, possibly leading to disclosure of sensitive information.

References: [1] https://www.kernel.org/pub/software/scm/git/docs/git-imap-send.html [2] https://github.com/git/git/blob/master/imap-send.c#L233

CVE-2016-8568

Read out-of-bounds in gitoidnfmt: https://github.com/libgit2/libgit2/issues/3936

CVE-2016-8569

DoS using a null pointer dereference in gitcommitmessage: https://github.com/libgit2/libgit2/issues/3937

Proposed patch:

https://github.com/libgit2/libgit2/pull/3956

Git is a fast, scalable, distributed revision control system with an unusually rich command set that provides both high-level operations and full access to internals. Git defines a line-based protocol that is used to exchange information between Git and Git credential helpers. Some ecosystems (most notably, .NET and node.js) interpret single Carriage Return characters as newlines, which renders the protections against CVE-2020-5260 incomplete for credential helpers that treat Carriage Returns in this way. This issue has been addressed in commit b01b9b8 which is included in release versions v2.48.1, v2.47.1, v2.46.3, v2.45.3, v2.44.3, v2.43.6, v2.42.4, v2.41.3, and v2.40.4. Users are advised to upgrade. Users unable to upgrade should avoid cloning from untrusted URLs, especially recursive clones.