GHSA-VJ3M-2G9H-VM4P

Multiple RCE vectors were found in Grav CMS. Three are critical, two are high.

1. Unsafe unserialize() in JobQueue, direct RCE gadget (Critical)

system/src/Grav/Common/Scheduler/JobQueue.php:465 calls unserialize(base64_decode(...)) without restricting allowed_classes. The Job class has call_user_func_array($this->command, $this->args) in its execution path, which is a direct gadget chain, inject a serialized Job with command = 'system' and args = ['whoami'].

The same codebase actually has a Serializable trait that correctly restricts classes, so this inconsistency stands out.

2. Unsafe unserialize() in FileCache, arbitrary class instantiation (Critical)

system/src/Grav/Framework/Cache/Adapter/FileCache.php:75 does unserialize($value, ['allowed_classes' => true]). That true allows instantiation of any class. If an attacker can write to the cache directory (via any file write primitive), they get object injection → RCE.

3. Unsafe unserialize() in Session (High)

system/src/Grav/Common/Session.php:116, same allowed_classes => true pattern on session data. Lower severity since session storage is typically more restricted.

4. Command injection in git clone (Critical)

system/src/Grav/Console/Cli/InstallCommand.php:150, only $this->destination uses escapeshellarg(). The $data['branch'], $data['url'], and $data['path'] variables go directly into the shell command without escaping. Admin-accessible via plugin/theme installation.

5. SSTI blocklist bypass (High)

system/src/Grav/Common/Security.php:267-286, cleanDangerousTwig() blocks twig_array_map and twig_array_filter but not twig_array_reduce. Also missing file_get_contents and fwrite from the dangerous function blocklist. An attacker who can inject Twig templates can bypass the security filter.

All five are independently exploitable. The unserialize issues are the most concerning since they don't require admin access if there's any file write primitive.

ProScan AppSec | proscan.one

Maintainer note, fix applied (2026-04-24)

Fixed in Grav core on the 2.0 branch: commit c66dfeb5f (items #1, #2, #3, #4) and commit 38685ac25 + c66dfeb5f (item #5), ships in 2.0.0-beta.2.

All five vectors addressed:

1. Scheduler\JobQueue unsafe unserialize, serialized_job now carries a sibling serialized_job_hmac signed with Security::getNonceKey(). reconstructJob refuses to unserialize an item whose HMAC is missing/mismatched and falls through to the safe structured-fields rebuild. A tampered queue file can no longer smuggle a forged Job for direct RCE via Job::exec → call_user_func_array.
   → system/src/Grav/Common/Scheduler/JobQueue.php

2. FileCache unsafe unserialize, same HMAC-integrity approach; see separate GHSA-gwfr-jfjf-92vv.
   → system/src/Grav/Framework/Cache/Adapter/FileCache.php

3. Session::getFlashObject unsafe unserialize, payload now wrapped in a v2|<hmac>|<serialized> envelope; legacy/forged envelopes return null instead of triggering unserialize.
   → system/src/Grav/Common/Session.php

4. InstallCommand git clone shell injection, branch, url, and path values read from user/.dependencies are now passed through escapeshellarg, with a -- separator before url/path to block option-injection (e.g. --upload-pack=evil).
   → system/src/Grav/Console/Cli/InstallCommand.php

5. SSTI blocklist bypass, twig_array_reduce (the specific name called out) plus twig_array_some and twig_array_every added to cleanDangerousTwig's CALLABLE_DANGEROUS_NAMES alongside the existing twig_array_map/filter. More importantly, the new Twig content sandbox in 2.0.0-beta.2 blocks this class of attack at a different layer, see the sandbox work in 38685ac25.
   → system/src/Grav/Common/Security.php

Tests:

• tests/unit/Grav/Common/Security/UnserializeIntegritySecurityTest.php, 8 cases covering JobQueue + Session HMAC integrity.
• tests/unit/Grav/Common/Security/FileCacheSecurityTest.php.
• tests/unit/Grav/Common/Security/CleanDangerousTwigTest.php, new twig_array_* entries in providerCallbackFunctions.

Untrusted input reaches a shell command, allowing arbitrary commands to run on the host. Typical impact: code execution in the application's environment.

Kodem intelligence

1. What is GHSA-VJ3M-2G9H-VM4P? GHSA-VJ3M-2G9H-VM4P is a critical-severity OS command injection vulnerability in getgrav/grav (composer), affecting versions < 2.0.0-beta.2. It is fixed in 2.0.0-beta.2. Untrusted input reaches a shell command, allowing arbitrary commands to run on the host.
2. Which versions of getgrav/grav are affected by GHSA-VJ3M-2G9H-VM4P? getgrav/grav (composer) versions < 2.0.0-beta.2 is affected.
3. Is there a fix for GHSA-VJ3M-2G9H-VM4P? Yes. GHSA-VJ3M-2G9H-VM4P is fixed in 2.0.0-beta.2. Upgrade to this version or later.
4. Is GHSA-VJ3M-2G9H-VM4P exploitable, and should I be worried? Whether GHSA-VJ3M-2G9H-VM4P is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
5. What actually determines whether GHSA-VJ3M-2G9H-VM4P is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
6. How do I fix GHSA-VJ3M-2G9H-VM4P? Upgrade getgrav/grav to 2.0.0-beta.2 or later.