CVE-2025-12732: AI component: Info Disclosure leaks sensitive data
MEDIUMAny WordPress site running WP Import plugin with an OpenAI API key configured is leaking that key to any authenticated user with Author-level access or above — no exploit sophistication required. Rotate your OpenAI API keys immediately, set spending limits in your OpenAI dashboard, and update or disable the plugin. Audit your WordPress user roster for unnecessary Author+ accounts.
Risk Assessment
The CVSS score of 4.3 undersells the business risk. Stolen OpenAI API keys unlock unbounded cost harvesting, unauthorized inference, and potential data access via the AI service — none of which show up in traditional vulnerability scoring. Organizations with high OpenAI API spend or sensitive data flowing through LLM pipelines face financial and operational exposure well beyond what Medium severity implies. Exploitability is trivial: any registered WordPress user at Author level or above can extract the key without technical skills.
Severity & Risk
Attack Surface
Recommended Action
7 steps-
Immediately revoke and reissue any OpenAI API keys configured in WP Import plugin — assume compromise if you cannot confirm no unauthorized Author+ access.
-
Update WP Import to a patched version above 7.33.
-
Set hard spending limits and usage alerts on your OpenAI account as a blast-radius control.
-
Review OpenAI API usage logs for anomalous call volumes or unexpected model usage.
-
Audit WordPress user accounts — remove or downgrade unnecessary Author+ roles.
-
Scope API keys to minimum required permissions and consider per-environment key rotation.
-
Detection: alert on OpenAI API calls from unexpected IP ranges or at unusual hours.
CISA SSVC Assessment
Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.
Classification
Compliance Impact
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2025-12732?
Any WordPress site running WP Import plugin with an OpenAI API key configured is leaking that key to any authenticated user with Author-level access or above — no exploit sophistication required. Rotate your OpenAI API keys immediately, set spending limits in your OpenAI dashboard, and update or disable the plugin. Audit your WordPress user roster for unnecessary Author+ accounts.
Is CVE-2025-12732 actively exploited?
No confirmed active exploitation of CVE-2025-12732 has been reported, but organizations should still patch proactively.
How to fix CVE-2025-12732?
1. Immediately revoke and reissue any OpenAI API keys configured in WP Import plugin — assume compromise if you cannot confirm no unauthorized Author+ access. 2. Update WP Import to a patched version above 7.33. 3. Set hard spending limits and usage alerts on your OpenAI account as a blast-radius control. 4. Review OpenAI API usage logs for anomalous call volumes or unexpected model usage. 5. Audit WordPress user accounts — remove or downgrade unnecessary Author+ roles. 6. Scope API keys to minimum required permissions and consider per-environment key rotation. 7. Detection: alert on OpenAI API calls from unexpected IP ranges or at unusual hours.
What systems are affected by CVE-2025-12732?
This vulnerability affects the following AI/ML architecture patterns: WordPress + LLM plugin integrations, CMS-embedded AI enrichment pipelines, Shared-credential API key configurations.
What is the CVSS score for CVE-2025-12732?
CVE-2025-12732 has a CVSS v3.1 base score of 4.3 (MEDIUM). The EPSS exploitation probability is 0.04%.
Technical Details
NVD Description
The WP Import – Ultimate CSV XML Importer for WordPress plugin for WordPress is vulnerable to unauthorized access of sensitive information due to a missing authorization check on the showsetting() function in all versions up to, and including, 7.33. This makes it possible for authenticated attackers, with Author-level access or higher, to extract sensitive information including OpenAI API keys configured through the plugin's admin interface.
Exploitation Scenario
An attacker registers a free account on a WordPress site (or compromises an existing low-privilege contributor account). They craft a direct HTTP request to the vulnerable showsetting() endpoint — no special tooling required, just an authenticated session cookie. The endpoint returns admin plugin configuration including the plaintext OpenAI API key. The attacker then uses this key externally to run GPT-4 or DALL-E queries at the victim's expense, or resells the key on underground markets. A financially motivated adversary could systematically scan WordPress sites running this plugin, extract keys at scale, and run a cost-harvesting operation before the victim notices unusual billing.
Weaknesses (CWE)
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N References
- plugins.trac.wordpress.org/browser/wp-ultimate-csv-importer/trunk/controllers/SendPassword.php
- plugins.trac.wordpress.org/browser/wp-ultimate-csv-importer/trunk/controllers/SendPassword.php
- plugins.trac.wordpress.org/changeset/3390161/wp-ultimate-csv-importer/trunk/controllers/SendPassword.php
- wordfence.com/threat-intel/vulnerabilities/id/25687ee6-a899-4089-966b-69578afd3fb6
Timeline
Related Vulnerabilities
CVE-2025-5120 10.0 smolagents: sandbox escape enables unauthenticated RCE
Same attack type: Data Leakage CVE-2025-53767 10.0 Azure OpenAI: SSRF EoP, no auth required (CVSS 10)
Same attack type: Data Extraction CVE-2023-3765 10.0 MLflow: path traversal allows arbitrary file read
Same attack type: Data Leakage CVE-2025-2828 10.0 LangChain RequestsToolkit: SSRF exposes cloud metadata
Same attack type: Data Extraction CVE-2026-21858 10.0 n8n: Input Validation flaw enables exploitation
Same attack type: Data Extraction
AI Threat Alert