CVE-2026-34106: Guardian Language-System Unauthenticated OS Command Injection via id Parameter in subtitles.php
Guardian language-system passes the id GET parameter directly into a PHP exec() call in subtitles.php (line 19) without sanitization: exec(\"php jobs/subtitle_rendering.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to the id parameter to execute arbitrary OS commands on the server.
Metrics
- CVSS v4.0
- 9.3
- Severity
- CRITICAL
- Fixed in
- —
- Affected Products
- 1
HarborGuard Analysis
Synopsis
An unauthenticated OS command injection vulnerability affects Guardian language-system (all versions through commit e42c395ec4b03fe62973a669c9209a673838b8a4). The flaw is reachable over the network with no credentials required: the subtitles.php script passes a user-supplied GET parameter directly into a PHP exec() call without any sanitization. A remote attacker can append shell metacharacters to the id parameter and execute arbitrary operating system commands on the server. No fix version has been published; HarborGuard tracks this advisory for patch availability.
HarborGuard Coverage
Detection of CVE-2026-34106 is available across every HarborGuard environment: the CVE is ingested from upstream feeds within minutes of publication and matched against all customer images, including custom-built images that bundle Guardian language-system. Any image containing an affected version of the package is flagged automatically.
AvailableHarborGuard is capable of scoring this CVE at its published severity (CVSS 9.3 Critical, v4.0) and weighting that score against each customer environment's compliance policy. Findings are routed to the appropriate team inbox within each customer organization based on policy configuration.
AvailableBecause no upstream fix has been published, HarborGuard re-checks this advisory on every ingest cycle and will make a patched-image rebuild available the moment a remediated version is released by the maintainer. In the interim, compensating controls such as network-policy isolation around services running Guardian language-system can be configured and enforced through HarborGuard's policy engine.
Pending upstreamExploit Conditions
- Network reachabilityRequired
The vulnerable subtitles.php endpoint is exposed over the network; an attacker must be able to send HTTP requests to the server to reach it.
- AuthenticationNot required
No credentials or session token of any kind are needed; the vulnerable code path is entirely unauthenticated.
- Victim interactionNot required
Exploitation is fully attacker-driven and requires no action from any user or administrator on the target system.
- Attack complexityDetail
Exploitation is reliable and condition-free; appending shell metacharacters to the id GET parameter is sufficient to inject commands with no timing dependencies or environmental prerequisites.
Blast Radius
- A successful attacker executes arbitrary OS commands as the web server process user, gaining an interactive foothold on the host.
- The attacker reads any file accessible to the web server process, including application source code, configuration files, and stored credentials.
- The attacker writes or overwrites files on the server, enabling webshell placement or modification of application data.
- The attacker can crash or abuse server processes, disrupting availability of the Guardian language-system service and any co-located workloads.
How HarborGuard Handles This
Available on HarborGuard: because no upstream fix exists for CVE-2026-34106, the platform continuously re-checks the advisory on every ingest cycle and will make a patched-image rebuild available automatically the moment the Guardian language-system maintainer publishes a remediated version. For customers who opt into auto-remediation, that rebuild will trigger a regression test run and a PR opened against affected workloads without manual intervention. While no patch is available, the recommended compensating controls are: restrict network access to the subtitles.php endpoint via Kubernetes NetworkPolicy or equivalent egress and ingress filtering rules; apply a web application firewall rule to block requests containing shell metacharacters in the id parameter; and, where the subtitle rendering feature is not required, gate it off via a feature flag or by removing the route entirely from the deployed image. HarborGuard's policy engine can surface images running the affected commit hash and flag them for immediate review in each customer environment's compliance dashboard.
- guardian / language-system≤ e42c395ec4b03fe62973a669c9209a673838b8a4
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N