Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessObtain WordPress Editor credentials
-
DeliveryAuthenticate to WordPress admin panel
-
ExploitNavigate to AI Engine configuration admin page
-
ExecutionInspect inline page JavaScript source
-
PersistExtract third-party API keys and tokens in cleartext
-
ImpactAbuse exfiltrated credentials against AI provider
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V14 - Data Protection Data exposure prevention
MITRE ATT&CK
2 techniques across 2 tactics: Collection, Discovery The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires a valid WordPress user account with at minimum the Editor role - unauthenticated remote attackers cannot exploit this vulnerability. The AI Engine plugin must be installed, active, and configured with at least one third-party API key or authentication token stored in the database; installations with no stored secrets expose nothing of value. The attacker must be able to reach the specific admin page where AI Engine renders its inline configuration script (standard WordPress admin panel network access). Exploitation is bounded by: (1) requiring authenticated Editor or higher access, which limits the attacker pool to insiders or accounts obtained through separate compromise; (2) the target site must have AI Engine actively configured with credentials; (3) the attack is not automatable per SSVC due to the authentication requirement. |
| Impact | The CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N) scores 6.5 (Medium), accurately reflecting network-accessible exploitation requiring only Editor-level credentials with no additional complexity. However, the EPSS score of 0.17% at the 7th percentile indicates very low real-world exploitation probability despite the public POC, suggesting opportunistic exploitation has not yet materialized at scale. SSVC assessment corroborates this: exploitation status is POC-only, the attack is not automatable (requires authenticated access with a specific role), and technical impact is partial. The actual business risk is disproportionate to the CVSS score in environments that store high-value credentials (e.g., paid OpenAI API keys): a compromised API key can result in financial abuse, unauthorized data access through the AI service, or lateral pivot if the same credentials are reused. Sites where no API keys are configured expose no sensitive data. Not currently in CISA KEV. |
| Exploit ScenarioDerived | An attacker with a legitimate WordPress Editor account - obtained via credential compromise, insider access, or social engineering - logs into the admin panel and navigates to the AI Engine configuration page. Without any special tooling, they view the page source or browser developer tools to locate the inline JavaScript block containing the plugin's configuration, where API keys and authentication tokens appear in cleartext. A public POC demonstrating this technique is available via WPScan, lowering the skill bar for exploitation. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Aug 06, 2026
Cleartext exposure of third-party API keys and authentication tokens in AI Engine WordPress plugin (before 3.6.4) allows Editor-role users to read secrets that are intended to be administrator-only. The plugin inlines sensitive configuration values into admin page JavaScript without redaction, making them readable to any authenticated user with Editor access who inspects the page source. A public proof-of-concept exists per WPScan, though EPSS remains low at 0.17% (7th percentile) and the vulnerability is not in the CISA KEV catalog.
RemediationAI
Update AI Engine to version 3.6.4 or later, which redacts sensitive configuration values before they are rendered into admin page inline scripts. The fix is confirmed available per vendor and documented in the WPScan advisory at https://wpscan.com/vulnerability/badce355-e6ce-4dd4-8d9b-c87a7b9db12f/ and NVD at https://nvd.nist.gov/vuln/detail/CVE-2026-16954. If immediate patching is not possible, a compensating control is to audit and reduce the number of accounts with the Editor role, or temporarily revoke Editor access from untrusted users - note this will disrupt normal editorial publishing workflows. Rotating all stored third-party API keys and authentication tokens is strongly advised if any Editor-role account is considered potentially hostile, since exposure may have already occurred on unpatched installations. Do not rely solely on network-level controls, as exploitation requires only legitimate authenticated access to the WordPress admin panel.
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
AI Engine is a WordPress plugin that integrates AI services such as OpenAI, storing third-party API credentials and authentication tokens in the WordPress database. CWE-200 (Exposure of Sensitive Information to an Unauthorized Actor) manifests here because the plugin serializes its full configuration - including secrets - directly into inline JavaScript within an admin page, without filtering or redacting sensitive fields before output. WordPress's role model grants Editor accounts access to portions of the admin interface, and this particular page is reachable at that privilege level, breaking the plugin's own access control boundary. The root cause is absent output sanitization in the server-side rendering path that builds the inline script block. Affected product is identified via CPE cpe:2.3:a:unknown:ai_engine:*:*:*:*:*:*:*:* covering all versions below 3.6.4.
Know when Ai Engine 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
AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N
Editor-role authentication required drives PR:L; full API key exposure in cleartext justifies C:H; no integrity or availability impact applies to this read-only disclosure.
Primary rating from Vendor (WPScan).
CVSS VectorVendor: WPScan
Lifecycle Timeline
3DescriptionCVE.org
The AI Engine WordPress plugin before 3.6.4 does not redact secret configuration values before exposing them in an admin page's inline script data, allowing users with the Editor role to read the site's stored third-party API key and authentication tokens in cleartext, despite those secrets being restricted to administrators everywhere else.
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
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.
Unrestricted file upload in Jordy Meow's AI Engine: ChatGPT Chatbot plugin for WordPress (versions up to and including 1
Arbitrary file write in the AI Engine WordPress plugin before 3.5.5 lets authenticated editor-level users abuse an unsan
Path traversal in the AI Engine WordPress plugin (versions 3.3.3 through 3.7.1) allows arbitrary server file reads with
Privilege escalation in the AI Engine WordPress plugin (versions 2.8.0 through 3.6.0) enables any sub-site Administrator
The AI Engine WordPress plugin before 2.6.5 does not sanitize and escape a parameter from one of its RESP API endpoint b
AI Engine < 2.4.3 is susceptible to remote-code-execution (RCE) via Log Poisoning. Rated high severity (CVSS 7.2), this
Server-Side Request Forgery (SSRF) vulnerability in Jordy Meow AI Engine: ChatGPT Chatbot.1.4. Rated medium severity (CV
Unauthorized access to private chatbot conversations in the AI Engine WordPress plugin before 3.5.5 allows any authentic
Unauthenticated remote attackers can invoke AI queries against a site owner's configured provider account through the AI
Unrestricted Upload of File with Dangerous Type vulnerability in Jordy Meow AI Engine: ChatGPT Chatbot.2.63. Rated criti
Unrestricted Upload of File with Dangerous Type vulnerability in Jordy Meow AI Engine: ChatGPT Chatbot.1.4. Rated critic
Path traversal in the AI Engine WordPress plugin before 3.6.6 allows server-side file contents to be read and exfiltrate
EUVD-2026-53726
GHSA-9wjr-j39x-f62m
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB