text-generation-webui vulnerabilities
4 known CVE vulnerabilities in text-generation-webui, translated and rated.
- CVE-2026-35487Medium
text-generation-webui prior to 4.3 contains an unauthenticated path traversal vulnerability in load_prompt() that allows reading any .txt file on the server filesystem. The file content is returned verbatim in the API response.
- CVE-2026-35484Medium
text-generation-webui before 4.3 has an unauthenticated path traversal vulnerability in load_preset() that allows reading any .yaml file on the server filesystem. Parsed YAML key-value pairs (including passwords, API keys, connection strings) are returned in the API response. Fixed in 4.3.
- CVE-2026-35483Medium
text-generation-webui prior to 4.3 has an unauthenticated path traversal vulnerability in load_template() allowing reading files with .jinja, .jinja2, .yaml, or .yml extensions from anywhere on the server filesystem. For .jinja files content is returned verbatim; for .yaml a parsed key is extracted. Fixed in 4.3.
- CVE-2025-62364Medium
In text-generation-webui versions through 3.13, a Local File Inclusion vulnerability exists in the character picture upload feature. An attacker can upload a text file containing a symbolic link to an arbitrary file path, allowing an unauthenticated attacker to read sensitive files on the server.
Vulnerability data from NVD (NIST) · CISA KEV · EPSS

