LibreChat exposed API keys in shared agent actions
- Exploited
- No known exploitation
- Fix
- Fixed in 0.8.7
- Who is exposed
- Not reported
Summary
LibreChat, an enhanced ChatGPT clone that supports multiple AI providers, lets a collaborator with edit permission on a shared agent, reached over the network after logging in with low privileges, repoint an existing action to an attacker-controlled host. When a tool later uses that action, LibreChat sends the owner's retained API key or OAuth client secret to that host, exposing the owner's credentials.
Written by a model
Original description
LibreChat is an enhanced ChatGPT clone that supports multiple AI providers. Prior to 0.8.7, POST /api/agents/actions/:agent_id in api/server/routes/agents/actions.js loads an existing action with owner credentials and merges editor-controlled metadata.domain and metadata.raw_spec while preserving metadata.api_key and related OAuth client secrets. A collaborator with EDIT permission can repoint an action to an attacker-controlled host, and a later tool invocation decrypts and sends the retained owner credential to that destination. This issue is fixed in version 0.8.7.
What to do
Upgrade to a fixed version: 0.8.7.
Check your systems with an agent
Paste this into a coding or ops agent. It checks your systems read-only and reports a table per asset.
Check whether my systems are affected by CVE-2026-54044 (LibreChat: Shared-Agent Editor Can Exfiltrate Owner Action API Keys via OpenAPI Spec Swap), using the cve.now data below.
Rules: work read-only. Do not run exploit code, proof-of-concept scripts or scanners that send attack payloads, and do not change any configuration, package, image or deployment. Report findings only.
What cve.now has for CVE-2026-54044 (entry last updated 2026-11-20):
Affected:
- LibreChat-AI LibreChat: fixed in 0.8.7
Fixed versions: 0.8.7.
Setup: unknown whether the bug needs a non-default feature or setting. Check the advisory.
CISA KEV: CVE-2026-54044 is not listed as of cve.now's last read.
This KEV status is for CVE-2026-54044 only. If you also check another CVE, look up its own KEV status; never assume it from this one.
How to check:
- Installed software: check what is actually running, not what a repo declares. Read the version from each host, container image and Kubernetes node (the binary's version output, the OS package database, image layers), and check whether the conditions above hold there. Distribution packages may carry a backported fix under an older version number; confirm against the distribution's own advisory.
Verify against the linked advisories before concluding, and cite the date you read each one:
- https://github.com/LibreChat-AI/LibreChat/security/advisories/GHSA-qh43-24gm-5j57
- https://github.com/LibreChat-AI/LibreChat/pull/13559
- https://github.com/LibreChat-AI/LibreChat/commit/07af6ee288b1fdb610105ffbac7c1f34103a8f46
- https://github.com/LibreChat-AI/LibreChat/releases/tag/v0.8.7
- https://www.librechat.ai/changelog/v0.8.7
Do not conclude an asset is unaffected unless you checked it and the evidence covers both its version and the conditions above. If you could not check something, mark it unknown and say what is missing.
Report a table with one row per asset and these columns: Asset | Version or build | Status (affected, fixed, not applicable, unknown) | Compensating controls | Exposure (internet-facing, internal, none) | Evidence | Next action.
Entry: https://cve.now/CVE-2026-54044
For follow-up questions, use the cve.now MCP server at https://mcp.cve.now/mcp (no authentication), for example execute with name cve.get and arguments {"id": "CVE-2026-54044"}.Affected versions
- LibreChat-AI LibreChatfixed in 0.8.7
cve.now Index
- Exploitation?/40Unknown. No CISA SSVC exploitation rating yet
- Exposure3/25Reachable over the network · Unknown: No advisory context yet
- Reach?/20Unknown. No tracked technology or package matched
- Severity7.7/10CVSS 7.7
- No patch0/5No signal
What would change the score
- Would score about 66 if added to CISA KEV
- Would score about 61 if CISA's SSVC rates exploitation active
- Would score about 30 if advisories confirm default setups are exposed
Timeline
- CVE published
- First seen by cve.now
Evidence 7
| Kind | Source |
|---|---|
| Vendor advisory 2Establishes: Affected and fixed versions. | |
| Advisory | GitHub · LibreChat-AI/LibreChat · Advisory GHSA-qh43-24gm-5j57 |
| Link | librechat.ai · changelog/v0.8.7 |
| CVE record 2Establishes: The identifier and description. | |
| Record | CVE List · CVE-2026-54044 |
| Record | NVD · CVE-2026-54044 |
| Other links 3 | |
| Repository | GitHub · LibreChat-AI/LibreChat · Pull request #13559 |
| Repository | GitHub · LibreChat-AI/LibreChat · Commit 07af6ee2 |
| Repository | GitHub · LibreChat-AI/LibreChat · Release v0.8.7 |
Questions about CVE-2026-54044
Is CVE-2026-54044 being exploited?
It is not in CISA's Known Exploited Vulnerabilities catalog, and CISA's SSVC assessment does not mark exploitation as active. That does not mean it has not been attacked: vendors and researchers can report attacks before CISA lists them. No public exploit code is tracked.
Source: cve.now feed data
Which versions are affected by CVE-2026-54044?
LibreChat: before 0.8.7.
Source: cve.now feed data
How do I fix CVE-2026-54044?
Upgrade to a fixed version: 0.8.7.
Source: cve.now feed data
Related
- CVE-2026-54026LibreChat: Cross-User Message Injection via Parameter Confusion
- CVE-2026-54038LibreChat: Server-Side ReDoS via Unescaped Regex in Principal Search
- CVE-2026-54031LibreChat: MCP OAuth requests can follow redirects to internal URLs after SSRF validation
- CVE-2026-54046LibreChat: Unauthenticated User Enumeration via Email Verification Endpoint
- CVE-2026-131287Remote Code Execution (RCE) through the REST API in Splunk SOAR
- CVE-2026-132659In Postgres Enterprise Manager (PEM), the row-level security policies on pem.job and pem.jobstep that apply to PEM agents do not limit an…
No corrections so far. Something wrong on this page? Report an error.