Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessLure victim to malicious page
-
DeliverySend localhost GET /better-bibtex/cayw with application=sublime
-
ExploitInject shell metacharacters in command param
-
ExecutionVictim completes CAYW selection dialog
-
Impactsh -c executes injected commands as JabRef user
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A05:2025 - Injection
ASVS 5.0
V1 - Encoding and Sanitization OS command sanitization, avoid shell exec
MITRE ATT&CK
2 techniques across 1 tactic: Execution The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires ALL of: (1) JabRef's built-in HTTP server enabled or jabsrv running - both disabled by default, so the operator must have opted in; (2) a valid Sublime Text command path configured as a push target in JabRef; (3) the ability to cause a localhost request to GET /better-bibtex/cayw with application=sublime (e.g. via a malicious web page or local process); and (4) the victim actively completing the CAYW citation selection dialog (UI:A). Only Unix-like systems are affected, since PushToSublimeText routes the command through sh -c. These layered non-default and interaction requirements are the primary limiting factors, matching the vendor's AC:H/AT:P/UI:A rating. |
| Impact | This is a genuine remote command-execution flaw (CWE-78) but with a stacked set of preconditions that sharply reduce real-world risk. The vendor CVSS 4.0 vector (AV:N/AC:H/AT:P/PR:N/UI:A/VC:H/VI:H/VA:H, 7.5) already encodes this: AC:H, AT:P (a specific attack requirement), and UI:A (active/passive victim interaction) all signal that the network-vector, no-privileges rating is heavily gated. Multiple conditions must line up simultaneously: the built-in HTTP server or jabsrv must be running (disabled by default), the victim must have configured a valid Sublime Text command path, and the victim must complete the CAYW selection dialog. No EPSS score, no CISA KEV listing, and no public exploit code were provided, so the exploitation status is 'no public exploit identified at time of analysis'. The impact when it does fire is severe - arbitrary OS command execution as the JabRef user - but the combination of a non-default feature, a specific editor configuration, and required user interaction makes this a lower-priority patch item than the raw C:H/I:H/A:H impact would suggest. Note the PR:N rating reflects that no JabRef authentication is needed to reach the localhost endpoint, not that the attack is condition-free. |
| Exploit ScenarioDerived | A victim who has enabled JabRef's built-in HTTP server (or runs jabsrv) and has configured Sublime Text as a push target visits an attacker-controlled web page; the page issues a localhost request to GET /better-bibtex/cayw with application=sublime and a command parameter containing shell metacharacters. When the victim completes the CAYW selection dialog, PushToSublimeText builds an unescaped string passed to sh -c via ProcessBuilder, executing the injected commands as the JabRef user. No public POC was provided in the intelligence set. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Sep 17, 2026
OS command injection in JabRef's built-in HTTP server (CWE-78) lets a caller who can reach the localhost GET /better-bibtex/cayw endpoint with application=sublime smuggle shell metacharacters through the untrusted cite-command prefix, which is concatenated with citation keys and executed via sh -c by PushToSublimeText without escaping, achieving arbitrary command execution as the JabRef user. The CVSS 4.0 vector (AV:N/AC:H/AT:P/PR:N/UI:A) shows this is remotely reachable and requires no authentication, but exploitation is gated behind a stack of non-default conditions: the built-in server must be enabled or jabsrv running, a valid Sublime Text command path must be configured as a push target, only Unix-like systems are affected because of the sh -c invocation, and the victim must actively complete the CAYW citation selection dialog. No KEV listing and no public exploit code were identified at time of analysis; the layered opt-in and interaction requirements keep real-world risk comparatively low despite the high confidentiality, integrity and availability impact if the preconditions are met. The flaw is fixed in 6.0-alpha.6.
RemediationAI
Upgrade to JabRef 6.0-alpha.6 or later, which is the vendor-released patch that fixes the CAYW command construction (upstream fix commit b8663fe58e6c87c3927cdb4eeb1f6934d7c3f1d2, PR #15628, release notes at https://github.com/JabRef/jabref/releases/tag/v6.0-alpha.6 and advisory GHSA-m42c-cw93-p629). If immediate upgrade is not possible, the most effective compensating control is to disable JabRef's built-in HTTP server in preferences and stop any running jabsrv instance - this removes the network-reachable attack surface entirely, at the cost of losing the CAYW citation-push integration with editors. Where the server must remain enabled, remove Sublime Text as a configured push target (the injection only fires when a valid Sublime Text command path is set) or disable the CAYW endpoint if a deployment-level toggle exists; note this breaks the cite-as-you-write workflow for Sublime Text users. As further hardening, restrict who can reach the server: bind it to loopback only and rely on OS-level controls so that arbitrary web content or non-privileged local processes cannot trigger the localhost request, since the endpoint is otherwise reachable by any browser on the machine. Because exploitation also depends on the victim completing the CAYW selection dialog, training users to treat unexpected citation-selection prompts as suspicious provides a last line of defense, though it is not a reliable technical mitigation. Both the server-disable and push-target-removal workarounds sacrifice functionality, so they should be considered temporary measures until the 6.0-alpha.6 upgrade can be applied.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
JabRef exposes a 'Cite As You Write' (CAYW) integration used by editors such as Sublime Text to pull citations from a running JabRef instance. When the built-in HTTP server (or the standalone jabsrv) is enabled, the GET /better-bibtex/cayw endpoint accepts an external 'command' query parameter. The root cause is CWE-78: CAYWQueryParams.getCommand() takes that attacker-controlled value and passes it through CAYWResource.getCitation() into PushToSublimeText.getCommandLine(), where the untrusted cite-command prefix and the citation keys are concatenated into a single command string. On Unix-like platforms, ProcessBuilder hands that string to sh -c without any shell escaping or argument separation, so shell metacharacters embedded in the parameter are interpreted by the shell rather than treated as literal text. The upstream fix commit (b8663fe58e6c87c3927cdb4eeb1f6934d7c3f1d2) and PR #15628 address the CAYW + Sublime Text and TeXShop command construction paths, and also route error dialogs in the push classes (GuiPushToSublimeText, GuiPushToEmacs, GuiPushToLyx, GuiPushToTeXstudio, GuiPushToTeXworks, GuiPushToTexShop, GuiPushToTexmaker, GuiPushToVScode) through UiTaskExecutor.runNowOrInJavaFXThread, indicating the fix hardens both the unsafe command building and the surrounding push workflow. Because the sink is a shell invocation rather than an exec-with-argv list, the Java-side process has the full privileges of the JabRef user, so post-exploitation impact is limited only by that account's permissions.
Know when Jabref gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
Reachable over network to a localhost endpoint with no JabRef auth (PR:N), but AC:H for the stacked non-default conditions and UI:R for the required CAYW dialog completion; full command execution gives C:H/I:H/A:H.
Primary rating from Vendor (github).
CVSS VectorVendor: github
Lifecycle Timeline
2DescriptionCVE.org
JabRef is a desktop application for managing BibTeX and BibLaTeX libraries. Prior to 6.0-alpha.6, when jabsrv or JabRef's built-in HTTP server is enabled, the GET /better-bibtex/cayw endpoint accepts an external command query parameter and CAYWQueryParams.getCommand() passes it through CAYWResource.getCitation() into PushToSublimeText.getCommandLine(). On Unix-like systems, PushToSublimeText combines this untrusted cite-command prefix and citation keys into a string executed through sh -c by ProcessBuilder without shell escaping. A client that can cause a localhost request with application=sublime can inject shell metacharacters and execute operating-system commands as the JabRef user when a valid Sublime Text command path is configured and the victim completes the CAYW selection dialog. The built-in server is disabled by default, so exploitation requires the victim to enable it or run jabsrv. This issue is fixed in version 6.0-alpha.6.
Technical POC Analysis
Note on evidence: the supplied material is the GitHub advisory GHSA-m42c-cw93-p629 plus the CVE summary.
Full technical POC analysis - root cause, step-by-step exploitation and patch breakdown - is part of the Pro and Business plans.
Executive ViewAI
JabRef’s optional local HTTP integration contains a high-severity command-injection flaw that, if the service is enabled and a user completes a citation-selection dialog, could let an unauthenticated attacker execute arbitrary commands with the user’s privileges. For research or engineering teams using JabRef with Better BibTeX and Sublime Text push configured, this creates risk of credential theft, data tampering, or lateral movement from an affected workstation. Leadership should prioritize inventory and temporary shutdown of the optional integration until a vendor patch is available.
No vendor-released patch identified at time of analysis. Interim mitigations include disabling the JabRef built-in HTTP server/jabsrv, disabling Better BibTeX CAYW and Sublime Text push targets, avoiding application=sublime, restricting the server to localhost and blocking external access, running JabRef as a low-privilege user, and monitoring shell/child-process activity and CAYW endpoint access.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
OS command injection occurs when an application passes unsanitized user input directly into system shell commands.
EUVD-2026-82565
NVD
CVE.org
EPSS
GitHub
Exploit-DB