CVE-2026-54024: LibreChat Unauthenticated File Upload DoS in Conversation Import

Technical summary

CVE-2026-54024 stems from inconsistent multer configuration across LibreChat's file upload routes. While commit bb58a2d0 addressed CVE-2024-11171 by adding limits: { fileSize } to the primary createMulterInstance() in file upload routes, the POST /api/convos/import endpoint instantiates a separate multer middleware instance that was never synchronized with these constraints. Additionally, the application-level validation controlled by CONVERSATION_IMPORT_MAX_FILE_SIZE_BYTES is commented out in the default .env.example, leaving no effective size enforcement on the import endpoint. An authenticated attacker can exploit this to conduct a denial-of-service attack by uploading files of unrestricted size, exhausting both disk I/O and memory resources. The vulnerability is classified as CWE-770 (Allocation of Resources Without Limits or Throttling).

Business impact

The vulnerability enables authenticated denial-of-service attacks that can render LibreChat instances unavailable. An internal user or compromised account could deplete server resources without triggering alerts, disrupting service for legitimate users. In multi-tenant or community-hosted deployments, this poses a particular risk: a single malicious user can degrade the platform for others. Recovery requires manual intervention (disk cleanup, service restart) and may result in data loss if the filesystem becomes full during critical operations. The attack requires valid authentication, reducing exposure in air-gapped environments but creating significant risk in cloud-hosted or public deployments where account compromise is a real threat vector.

Affected systems

LibreChat versions prior to 0.8.4-rc1 are affected. The vulnerability is specific to the POST /api/convos/import endpoint and requires authentication to exploit. Any deployment with user account provisioning and conversation import functionality enabled is in scope. The default configuration (with CONVERSATION_IMPORT_MAX_FILE_SIZE_BYTES commented out in .env.example) means most installations are vulnerable out of the box unless operators explicitly configured size limits after deployment.

Exploitability

Exploitability is moderate. The attack requires a valid authenticated session, which is a meaningful barrier compared to unauthenticated attacks, but not insurmountable. An insider, social-engineered user, or attacker with credentials from a separate compromise can readily abuse this. The POST /api/convos/import endpoint is likely discoverable through API documentation or client-side code inspection. No special tooling is needed—standard HTTP clients can perform the attack. The fact that the application-level check is disabled by default means most installations present no additional obstacles once authentication is obtained. The CVSS 6.5 (MEDIUM) score reflects the authentication requirement; without it, severity would be significantly higher.

Remediation

Upgrade to LibreChat 0.8.4-rc1 or later, which unifies file upload size enforcement across all endpoints. Organizations unable to upgrade immediately should enforce the CONVERSATION_IMPORT_MAX_FILE_SIZE_BYTES environment variable in their .env configuration to a reasonable value (e.g., 50MB) and monitor disk usage for anomalous growth. Consider implementing rate-limiting on the /api/convos/import endpoint and restrict access to trusted user groups if business logic permits. Review authentication logs for suspicious upload activity, particularly large or repeated import requests from single accounts.

Patch guidance

Apply the upgrade to 0.8.4-rc1 at your earliest convenience. Before patching, verify that your current configuration includes CONVERSATION_IMPORT_MAX_FILE_SIZE_BYTES set to an appropriate limit in your production .env file. If it is commented out or unset, add this configuration with a reasonable upper bound (consult your typical conversation file sizes) to mitigate risk immediately. Test the patch in a staging environment to confirm import functionality still works as expected with the new constraints. The fix is expected to be straightforward and low-risk, as it only adds constraints that were already present in other code paths; however, verify against the official vendor release notes to confirm no breaking changes to legitimate import workflows.

Detection guidance

Monitor HTTP request logs for POST /api/convos/import requests with unusually large Content-Length headers or multipart upload chunks. Alert on any single import request exceeding your expected file size threshold (e.g., >100MB). Track disk usage growth correlated with import endpoint activity; a sudden spike in disk I/O or filesystem utilization coinciding with import requests warrants investigation. Review authentication audit logs to identify which users are performing imports and cross-reference against expected behavior. If available, enable verbose multer or Express middleware logging to capture actual file sizes being processed on the import endpoint before the patch is applied.

Why prioritize this

This vulnerability merits prompt patching due to its denial-of-service impact and the fact that most installations are vulnerable by default. While authentication is required, it is not a barrier for internal threats or credential compromise. Patch priority depends on your threat model: if LibreChat is internal-only with strong identity controls, this is moderate priority; if it is user-facing or cloud-hosted with public or semi-public account registration, it is high priority. The presence of an easy environmental workaround (setting CONVERSATION_IMPORT_MAX_FILE_SIZE_BYTES) means you can reduce immediate risk without upgrading, but the permanent fix should not be deferred.

Risk score, explained

The CVSS 6.5 (MEDIUM) score reflects a network-accessible service (AV:N) with low attack complexity (AC:L) that requires authentication (PR:L) but causes high availability impact (A:H) with no confidentiality or integrity loss (C:N, I:N). The authentication requirement prevents this from being a 9+ severity issue; however, the high availability impact and ease of exploitation once authenticated justify the medium rating. In practice, organizations should weigh the environmental risk: public LibreChat instances with weak access controls may warrant internal high-severity treatment despite the CVSS medium designation.

Frequently asked questions