GHSA-VJP4-M83C-MP26
Vulnerability from github – Published: 2026-09-07 09:31 – Updated: 2026-09-08 18:31The ftp and scp tasks of Apache Ant can download files from a remote server. A malicious server can provide relative paths that allow it to write outside of the dedicated target directory for the download, making it possible to overwrite files of the attacker's choice using the permissions of the user running Ant in versions prior to Ant 1.10.18.
In order to exploit this vulnerability, the server would either have to be malicious or be subject to a machine-in-the-middle attack. Additionally in the case of scp or the ftp task using ftps the server must pass the server identity checks performed by the tasks.
For ftp tasks not using ftps a malicious server could act as a machine-in-the-middle to provide malicious files.
Starting with Ant 1.10.18 both tasks will prevent writing outside of the destination directory by default. An option is available to disable this behavior in the unlikely case that the old behavior is required by existing build files.
Mitigations:
Users of scp and ftp (when using ftps) in any version of Ant should not bypass server identity checks. Users of ftp not using ftps should switch to ftps where possible.
All users are recommended to upgrade to Apache Ant 1.10.18, which fixes this issue.
{
"affected": [],
"aliases": [
"CVE-2026-78254"
],
"database_specific": {
"cwe_ids": [
"CWE-23"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-09-07T08:17:12Z",
"severity": "HIGH"
},
"details": "The ftp and scp tasks of Apache Ant can download files from a remote server. A malicious server can provide relative paths that allow it to write outside of the dedicated target directory for the download, making it possible to overwrite files of the attacker\u0027s choice using the permissions of the user running Ant in versions prior to Ant 1.10.18. \n\n\n\n\nIn order to exploit this vulnerability, the server would either have to be malicious\u00a0or be subject to a machine-in-the-middle attack. Additionally\u00a0in the case of scp or the ftp task using ftps the server must pass the server identity checks performed by the tasks.\n\n\n\n\nFor ftp tasks not using ftps a malicious server could act as a machine-in-the-middle to provide malicious files.\n\n\n\n\nStarting with Ant 1.10.18 both tasks will prevent writing outside of the destination directory by default. An option is available to disable this behavior in the unlikely case that the old behavior is required by existing build files.\n\n\n\n\nMitigations:\n\n\n\n\nUsers of scp and ftp (when using ftps) in any version of Ant should not bypass server identity checks. Users of ftp not using ftps should switch to ftps where possible.\n\n\n\n\nAll users are recommended to upgrade to Apache Ant 1.10.18, which fixes this issue.",
"id": "GHSA-vjp4-m83c-mp26",
"modified": "2026-09-08T18:31:45Z",
"published": "2026-09-07T09:31:39Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-78254"
},
{
"type": "WEB",
"url": "https://lists.apache.org/thread/05s46qrlbhd06c6rsgvxh8zo0l4p9scz"
},
{
"type": "WEB",
"url": "http://www.openwall.com/lists/oss-security/2026/09/06/2"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:N",
"type": "CVSS_V3"
}
]
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.