PHP Fixes 7 Critical Flaws, Including Bugs That Could Allow Remote Code Execution
The fixes appear in PHP 8.2.36, 8.3.37, 8.4.28, 8.5.13, and 8.6.0. Not every vulnerability is fixed in every listed branch: for example, CVE-2026-129182 and CVE-2026-129474 are listed as fixed only in 8.5.13 and 8.6.0. This matters because administrators need to match the relevant fix to their PHP version and exposure. The article says the flaws affect all supported PHP versions under specific conditions. Those conditions can involve non-default configurations or uncommon setups. Some examples are quite specific: one flaw needs attacker influence over session data and a suitable deserialization gadget chain; another requires persistent backend connections and an application path that does not fully read php://input. The PHP Foundation released the security updates after maintainers issued an advance notice. The listed fixed versions give deployment teams concrete targets, but the article does not describe a single configuration that makes every deployment vulnerable. Checking the individual CVE conditions and applying the corresponding update are therefore important parts of assessing exposure.
Which PHP versions include the fixes, and what kinds of deployments may be affected?
The fixes appear in PHP 8.2.36, 8.3.37, 8.4.28, 8.5.13, and 8.6.0. Not every vulnerability is fixed in every listed branch: for example, CVE-2026-129182 and CVE-2026-129474 are listed as fixed only in 8.5.13 and 8.6.0. This matters because administrators need to match the relevant fix to their PHP version and exposure.
The article says the flaws affect all supported PHP versions under specific conditions. Those conditions can involve non-default configurations or uncommon setups. Some examples are quite specific: one flaw needs attacker influence over session data and a suitable deserialization gadget chain; another requires persistent backend connections and an application path that does not fully read php://input.
The PHP Foundation released the security updates after maintainers issued an advance notice. The listed fixed versions give deployment teams concrete targets, but the article does not describe a single configuration that makes every deployment vulnerable. Checking the individual CVE conditions and applying the corresponding update are therefore important parts of assessing exposure.
How many vulnerabilities are listed, and how do their severity scores compare?
The list contains eight vulnerabilities. Seven have a CVSS score of 9.2, while CVE-2026-129291 has a score of 8.8. The scores indicate that the list is dominated by flaws the article rates at 9.2, with one vulnerability receiving a somewhat lower score.
The vulnerabilities are not identical in type or effect. They include improper input validation, a time-of-check time-of-use issue, inconsistent interpretation of HTTP requests, and several use-after-free flaws. Reported consequences range from denial-of-service and session corruption to memory corruption, local privilege escalation, and remote code execution. Some have additional prerequisites for exploitation.
The high scores do not mean every PHP installation is automatically exploitable. The article notes that specific conditions apply, and some flaws affect only particular PHP versions or deployment arrangements. The release provides fixes across listed PHP branches, though the exact fixed versions differ by vulnerability. Readers should consider both the score and the stated conditions when understanding the risk.
What could happen if an attacker exploits these flaws, especially one that enables remote code execution?
The reported consequences include remote code execution, session corruption, denial-of-service, memory corruption, and local privilege escalation. Remote code execution is especially serious because the article describes attackers running code as the PHP worker or with its privileges. That can put the affected application process under an attacker’s control.
For example, CVE-2026-129289 could allow attacker-controlled PHP code to run as the worker when a front end keeps backend connections open and an application path opens php://input without fully reading it. Other listed flaws can crash workers or corrupt memory. CVE-2026-126714 can turn an interrupted session-file update into an arbitrary deserialization primitive that may be chained to code execution.
These outcomes are conditional, not guaranteed for every installation. Several flaws require particular application behavior, attacker influence over data, or specific server settings. The PHP Foundation has released fixes for the listed vulnerabilities, with fixed versions varying by flaw. Updating the affected PHP branch addresses the documented software defects; the article does not claim that exploitation has occurred on any particular server.
What is remote code execution, and why does it let an attacker affect a server?
Remote code execution, commonly shortened to RCE, means an attacker can cause a program on a remote system to execute code chosen or controlled by the attacker. This is a general security definition; the article does not define the term. RCE matters because it crosses the boundary from sending data to influencing what the server runs.
In this PHP report, the potential execution is described as code running with the privileges of a PHP worker. For CVE-2026-126714, the article says an attacker may manipulate session data and exploit an interrupted write that leaves a malformed session file. PHP may later deserialize that data, and a suitable gadget chain can lead to RCE.
The worker’s privileges determine what the code can do within that process and deployment; the article does not claim that every flaw grants unrestricted control of an entire server. Exploitation conditions differ across vulnerabilities. The listed updates fix the defects in specified PHP releases, so the real-world risk depends on both the affected flaw and the deployment’s configuration and application behavior.
How can a time-of-check time-of-use flaw and an interrupted session-file write lead PHP to read a corrupted session?
A time-of-check time-of-use, or TOCTOU, flaw occurs when a condition is checked and the relevant state changes before the program uses it. The article identifies CVE-2026-126714 as a TOCTOU flaw and says exploitation requires attacker influence over session data, a useful interruption during the write, and a suitable deserialization gadget chain.
The described mechanism involves PHP writing session data to disk. If a pwrite() or ftruncate() update is interrupted, a worker may die while the write is underway, or after it finishes but before old session content is removed. Either case can leave a file containing part of the new serialized session and leftover bytes from the previous one. PHP later loads it with session_start().
An attacker who can shape the new data may choose length fields so the session decoder interprets leftover bytes unexpectedly. The article says this can create an arbitrary deserialization primitive, which can be chained with another PHP-native finding to reach remote code execution. The fix is listed in PHP 8.2.36, 8.3.37, 8.4.28, 8.5.13, and 8.6.0.
What is a PHP session, and how can serialized session data be read back into objects by an application?
A PHP session is a way for an application to retain data associated with a session across separate requests. This is general PHP background, not a definition supplied in the article. The article focuses on session data stored in files and on PHP loading that data when session_start() runs.
Serialization represents data in a form that can be stored or transmitted. In PHP, session handling can encode values and decode them when a session is loaded; depending on the stored values and application, decoded data can include objects. The article describes CVE-2026-126714 as allowing a malformed session file to yield an arbitrary deserialization primitive. It does not provide a general tutorial on PHP object deserialization.
The reported danger is that an interrupted file update may leave a mix of new session data and bytes from an older session. PHP may later read that mixed file as session content. The article says a suitable deserialization gadget chain can turn the unexpected state into remote code execution. Exploitation requires attacker influence over session data and an interruption at a useful point, and the listed PHP updates fix the flaw.
What does “use after free” mean in memory management, and how can it cause crashes or code execution?
A use-after-free is a memory-management error: a program releases a block of memory but later tries to use it as if it were still valid. This is general computer-security background; the article names several PHP vulnerabilities as use-after-free flaws but does not explain the term.
The consequences depend on what happens after the memory is freed and how the program uses it again. The article reports that CVE-2026-129182 could cause memory corruption and possible full remote code execution. CVE-2026-129496 could let an unauthenticated remote attacker crash PHP workers or execute code with their privileges. Another flaw, CVE-2026-129481, could disrupt session cleanup, corrupt memory, or crash a reused worker.
These examples show why a memory-lifetime mistake can threaten both availability and security, though exploitation is not automatic. CVE-2026-129474, for example, affects PHP 8.5 and later and depends on application-specific conditions. The PHP Foundation lists fixes for these flaws in particular PHP releases; which update applies depends on the vulnerability and version branch.
This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.
Read more in the JupiteX app
Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.
Or read more news on the web