Skip to main content
Vulnerability Database/CVE-2026-108609

CVE-2026-108609: JeecgBoot IDOR Information Disclosure Flaw

CVE-2026-108609 is an insecure direct object reference flaw in JeecgBoot that exposes user AI voice generation history to unauthorized access. This post explains its impact, affected versions, and mitigation steps.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-108609 Overview

CVE-2026-108609 is an Insecure Direct Object Reference (IDOR) vulnerability [CWE-639] in JeecgBoot through version 3.9.5. The flaw exists in the GET /airag/voice/listByUser endpoint of the AI RAG voice module. Authenticated users can supply an arbitrary userId parameter to retrieve another user's AI voice generation history stored in Redis. Exposed data includes submitted text-to-speech input, voice settings, timestamps, and generated audio file names and paths.

Critical Impact

Any authenticated low-privilege user who knows or guesses another user's id can read that user's complete AI voice generation history, exposing potentially sensitive text input and file paths.

Affected Products

  • JeecgBoot through version 3.9.5
  • The jeecg-boot-module-airag voice module
  • Deployments exposing the /airag/voice/listByUser endpoint

Discovery Timeline

  • 2026-10-10 - CVE-2026-108609 published to the National Vulnerability Database (NVD)
  • 2026-10-13 - Last updated in NVD database

Technical Details for CVE-2026-108609

Vulnerability Analysis

The vulnerability resides in the VoiceController class of the jeecg-boot-module-airag module. The listByUser handler accepts a userId request parameter and returns the voice history records associated with that identifier from Redis. The controller does not verify that the authenticated session's user identity matches the supplied userId value. As a result, horizontal privilege escalation occurs across any pair of authenticated accounts.

Attackers who iterate or enumerate user identifiers can harvest text-to-speech prompts, voice configuration parameters, submission timestamps, and the names and paths of generated audio artifacts. Because the data stems from AI voice synthesis workflows, exposed prompts may contain confidential business content, personal messages, or voice-cloning source material.

Root Cause

The root cause is missing object-level authorization on the listByUser endpoint. The implementation treats the client-supplied userId parameter as authoritative instead of binding retrieval to the authenticated principal from the session or JWT. This matches the [CWE-639] pattern of authorization based on a user-controlled key.

Attack Vector

Exploitation requires only a valid authenticated session and network access to the application. The attacker issues an HTTP GET request to /airag/voice/listByUser with the userId query parameter set to a target user identifier. The server returns the matching Redis-backed history without ownership checks. A published proof-of-concept script demonstrates the request flow against a vulnerable instance.

See the GitHub PoC for CVE Request and the GitHub JeecgBoot Voice Controller Code for the vulnerable handler.

Detection Methods for CVE-2026-108609

Indicators of Compromise

  • Repeated GET requests to /airag/voice/listByUser where the userId parameter does not match the authenticated session's user identifier.
  • A single authenticated session enumerating many distinct userId values in a short time window.
  • Access patterns from accounts that have never previously invoked the airag/voice endpoints followed by sudden broad enumeration.

Detection Strategies

  • Instrument the VoiceController.listByUser handler to log the authenticated principal alongside the requested userId and alert on mismatches.
  • Add a reverse proxy or WAF rule that compares the session identity to the userId query parameter for all /airag/voice/* routes.
  • Correlate application access logs with Redis read patterns on voice history keys to surface cross-tenant retrieval.

Monitoring Recommendations

  • Baseline normal userId values submitted by each authenticated account and alert on deviations.
  • Monitor response sizes for /airag/voice/listByUser to detect bulk harvesting attempts.
  • Review the VulnCheck IDOR Advisory for JeecgBoot for additional detection guidance.

How to Mitigate CVE-2026-108609

Immediate Actions Required

  • Restrict or disable the /airag/voice/listByUser endpoint until a vendor patch is applied.
  • Enforce object-level authorization at a reverse proxy or API gateway by rejecting requests where the userId parameter differs from the authenticated user.
  • Rotate or purge any Redis-stored voice history records that may contain sensitive text-to-speech input.

Patch Information

As of publication, no vendor-released fixed version is referenced in the NVD entry for JeecgBoot. Organizations should monitor the JeecgBoot repository and the VulnCheck IDOR Advisory for JeecgBoot for an upstream patch. The appropriate code-level fix is to derive the lookup key from the authenticated principal rather than the client-supplied userId request parameter.

Workarounds

  • Deploy a middleware filter that overrides the userId parameter with the authenticated user's id before the controller executes.
  • Place the airag/voice routes behind an authorization policy that denies cross-user access by default.
  • Limit the AI voice feature to a trusted subset of users through role-based access control until a patched release is available.
bash
# Example nginx snippet to block cross-user requests to the vulnerable endpoint
location /airag/voice/listByUser {
    # Require the userId query param to equal the authenticated user id
    # supplied by an upstream auth service via the X-User-Id header
    if ($arg_userId != $http_x_user_id) {
        return 403;
    }
    proxy_pass http://jeecgboot_upstream;
}

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.