CVE-2026-13503: ANTLR4 Path Traversal in TokenVocab Grammar Handler

Technical summary

The vulnerability exists in the TokenVocabParser.java component within ANTLR4's grammar option handler, specifically in the getImportedVocabFile function. The flaw is a classic CWE-22 path traversal issue where user-controlled input from the tokenVocab grammar option is not properly sanitized before being used in file operations. This allows an attacker to use directory traversal sequences (such as '../') to escape the expected vocabulary file directory and access arbitrary files readable by the process. The vulnerability affects ANTLR4 versions up to and including 4.13.2.

Business impact

Organizations using ANTLR4 in production toolchains—particularly those processing untrusted grammar files or exposing grammar compilation as a service—face information disclosure risks. If ANTLR4 runs with elevated privileges or has access to sensitive configuration files, credentials, or source code, an attacker could extract that data. Development environments are particularly at risk if developers routinely process grammar files from external sources (third-party libraries, open-source projects, or user submissions). The impact is less severe than remote code execution but sufficient to compromise confidentiality of stored data.

Affected systems

ANTLR4 versions up to 4.13.2 are affected. Any application or development environment using ANTLR4 to parse or compile grammar files is in scope. This includes IDEs with ANTLR plugins, build systems that invoke ANTLR4 for code generation, and language-processing tools built on ANTLR4. The vulnerability is most dangerous in environments where grammar files come from untrusted or external sources.

Exploitability

Exploitability is straightforward: an attacker crafts a grammar file containing a malicious tokenVocab option with path traversal sequences and provides it to a system running vulnerable ANTLR4. No authentication or special privileges are required on the attacker side. Since exploits are now public, the barrier to weaponization is low. However, successful exploitation requires that the target system actually process the malicious grammar file—organizations that do not use ANTLR4 or do not accept untrusted grammar inputs are unaffected.

Remediation

Upgrade ANTLR4 to a patched version released after 4.13.2. Verify against the official ANTLR project releases and security advisories for the specific patched version number. As an interim control, restrict which users can invoke ANTLR4 grammar compilation, limit the grammar files processed to trusted sources only, and run ANTLR4 processes with minimal file system permissions and under a low-privilege service account to reduce the blast radius of file disclosure.

Patch guidance

Check the ANTLR4 GitHub repository and official release notes for versions newer than 4.13.2. The fix should address the path traversal in TokenVocabParser.java by properly validating or sanitizing the tokenVocab input before file operations. Apply the patch to all development systems, build pipelines, and any production environments that invoke ANTLR4. Test grammar processing workflows post-patch to ensure no regressions. Document the patched version in your software bill of materials.

Detection guidance

Monitor ANTLR4 process execution for attempts to access files outside expected directories, particularly in vocabulary file lookups. If available, enable ANTLR4 debug logging to capture grammar option parsing. In development environments, code review grammar files for suspicious tokenVocab values containing '../' or other path traversal patterns before processing. File system auditing (via tools like auditd on Linux or Sysmon on Windows) can flag unexpected file reads from ANTLR4 processes. Watch for error messages or exceptions related to file not found when processing grammar files, as these may indicate traversal attempts.

Why prioritize this

Although the CVSS score (5.3 MEDIUM) reflects limited direct impact, this vulnerability warrants prompt attention because: (1) exploits are publicly available; (2) the attack surface is any organization processing untrusted grammar files; (3) information disclosure can expose secrets or intellectual property; and (4) organizations may not immediately realize they're affected if ANTLR4 is a transitive dependency or embedded in a larger tool. Prioritize patching in development and CI/CD environments first, where untrusted grammar files are most likely to be processed.

Risk score, explained

The CVSS 3.1 score of 5.3 (MEDIUM) reflects a remotely exploitable vulnerability with low complexity and no privilege requirement, but limited impact scope—only confidentiality is affected (the 'C' in the vector is set, while 'I' and 'A' are not). The score does not account for public exploit availability or vendor non-responsiveness, which elevate practical risk. Organizations with tight control over grammar file sources may assess lower risk; those accepting external grammars or exposing grammar compilation services should treat this as higher priority.

Frequently asked questions