{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 0.8.5"
      },
      "package": {
        "ecosystem": "PyPI",
        "name": "open-webui"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.8.6"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-45350"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-862"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-05-14T20:24:48Z",
    "nvd_published_at": "2026-05-15T22:16:55Z",
    "severity": "HIGH"
  },
  "details": "### Summary\nOpen WebUI v0.6.43 contains a vulnerability in its chat completion API, which allows attackers to bypass tool restrictions, potentially enabling unauthorized actions or access.\n\n\n### Details\nIn the [chat_completion](https://github.com/open-webui/open-webui/blob/a7271532f8a38da46785afcaa7e65f9a45e7d753/backend/open_webui/main.py#L1529) API, the parameters [tool_ids and tool_servers](https://github.com/open-webui/open-webui/blob/a7271532f8a38da46785afcaa7e65f9a45e7d753/backend/open_webui/main.py#L1613-L1614) are supplied by the user. These parameters are used to [create a tools_dict by the middleware](https://github.com/open-webui/open-webui/blob/a7271532f8a38da46785afcaa7e65f9a45e7d753/backend/open_webui/utils/middleware.py#L1394). This is then used by [get_tool_by_id](https://github.com/open-webui/open-webui/blob/a7271532f8a38da46785afcaa7e65f9a45e7d753/backend/open_webui/models/tools.py#L139) to retrieve the appropriate tool. However, there is no checks in that ensures the user that uses the API has permission to use the tool, meaning that a user can invoke any server tool by supplying the correct tool_id or tool_servers parameters via the chat completion API. Moreover, the authentication token stored in the server would be used when invoking the tool, so the tool will be invoked with the server privilege.\n### PoC\nTo reproduce the issue, create an admin user and create an external tool via the admin settings. Set the type to \"MCP Streamable HTTP\" for the tool, and the url pointing to a mcp server. For example, the public instance of the [\"Fetch\" mcp server](https://github.com/modelcontextprotocol/servers/blob/main/src/fetch/README.md) can be used, which is located at \"https://remote.mcpservers.org/fetch/mcp\". Other mcp servers, e.g. the GitHub MCP server can also be used. Then set the \"Auth\" field appropriately. (Fetch mcp does not require Auth to be set).\n\nSet the ID, name and description for the MCP server and set visibility to private. This should prevent any user from using the mcp server. (For the example below, we use the fetch mcp server and set the ID to 1)\n\nNext create a user with low privilege and enable API keys from the admin settings.\n\nThen use the chat completion API with the low privilege user with a prompt specifying the use of the restricted tool, and include the id of the tool in the tool_ids parameter. For example, to use the fetch mcp server set up before:\n\n```\ndef chat_with_model(token):\n    url = \u0027http://localhost:3000/api/chat/completions\u0027\n    headers = {\n        \u0027Authorization\u0027: f\u0027Bearer {token}\u0027,\n        \u0027Content-Type\u0027: \u0027application/json\u0027\n    }\n    data = {\n      \"model\": \"llama3.1:latest\",\n      \"messages\": [\n        {\n          \"role\": \"user\",\n          \"content\": \"Use the fetch tool to fetch content of the url https://raw.githubusercontent.com/modelcontextprotocol/servers/refs/heads/main/src/fetch/LICENSE\"\n        }],\n        \n        \"tool_ids\" : [\n            \"server:mcp:1\",\n        ],\n    }\n    response = requests.post(url, headers=headers, json=data)\n    return response.json()\n```\n\nNote that the tool will be used to fetch the content of the file, despite the tool is restricted and has it\u0027s visibility set to private\n\n### Impact\nThis issue may lead to restricted tools being invoked by users via the chat completion API\n\n## Resolution\n\nFixed across two releases\n\n- **Local Tool records** (`tool_ids: [\"\u003ctool_id\u003e\"]` referencing a stored Tool): fixed in commit [9b06fdc8f](https://github.com/open-webui/open-webui/commit/9b06fdc8fe1c933071610336be05f11e77e6c8eb), first released in **v0.7.0**. `get_tools()` in `backend/open_webui/utils/tools.py` (line 166) now resolves the caller\u0027s group memberships and rejects each requested `tool_id` whose owner isn\u0027t the caller and for which no `AccessGrants.has_access(resource_type=\u0027tool\u0027, permission=\u0027read\u0027)` grant exists. Admins continue to bypass when `BYPASS_ADMIN_ACCESS_CONTROL` is enabled (its documented UI/posture purpose).\n\n- **Admin-configured MCP servers** (`tool_ids: [\"server:mcp:\u003cid\u003e\"]`, the report\u0027s exact PoC): fixed in commit [4737e1f11](https://github.com/open-webui/open-webui/commit/4737e1f11), first released in **v0.8.6**. The MCP-resolution loop in `backend/open_webui/utils/middleware.py` (line 2670) now calls `has_connection_access(user, mcp_server_connection)` and skips the server with a warning if the user has no grant \u2014 the server\u0027s stored credentials are never used on behalf of an unauthorized caller.\n\nUsers on `\u003e= 0.8.6` are not affected.",
  "id": "GHSA-4pcg-253r-rf9w",
  "modified": "2026-05-19T15:59:53Z",
  "published": "2026-05-14T20:24:48Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/open-webui/open-webui/security/advisories/GHSA-4pcg-253r-rf9w"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-45350"
    },
    {
      "type": "WEB",
      "url": "https://github.com/open-webui/open-webui/commit/4737e1f11"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/open-webui/open-webui"
    },
    {
      "type": "WEB",
      "url": "https://github.com/open-webui/open-webui/releases/tag/v0.8.6"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Open WebUI\u0027s chat completion API allows tool restrictions to be bypassed"
}