Common Weakness Enumeration

CWE-639

Allowed

Authorization Bypass Through User-Controlled Key

Abstraction: Base · Status: Incomplete

API response cwe.mitre.org

The system's authorization functionality does not prevent one user from gaining access to another user's data or record by modifying the key value identifying the data.

3819 vulnerabilities reference this CWE, most recent first.

GHSA-QVQR-5CV7-WH35

Vulnerability from github – Published: 2026-03-27 18:36 – Updated: 2026-03-30 20:10
VLAI
Summary
MCP Ruby SDK: Insufficient Session Binding Allows SSE Stream Hijacking via Session ID Replay
Details

Summary

The Ruby SDK's streamable_http_transport.rb implementation contains a session hijacking vulnerability. An attacker who obtains a valid session ID can completely hijack the victim's Server-Sent Events (SSE) stream and intercept all real-time data.

Details

Root Cause The StreamableHTTPTransport implementation stores only one SSE stream object per session ID and lacks:

  • Session-to-user identity binding
  • Ownership validation when establishing SSE connections
  • Protection against multiple simultaneous connections to the same session

PoC

Vulnerable Code

File: streamable_http_transport.rb - L336-L339:

def store_stream_for_session(session_id, stream)
  @mutex.synchronize do
    if @sessions[session_id]
      @sessions[session_id][:stream] = stream  # OVERWRITES existing stream
    else
      stream.close
    end
  end
end

Attack Scenario

Step 1: Legitimate Session Establishment

POST / (initialize) → receives session_id: "abc123"
GET / with Mcp-Session-Id: abc123 → SSE stream connected

Step 2: Session ID Compromise

  • An attacker obtains the session ID through various means (out of scope for this analysis)

Step 3: Stream Hijacking

GET / with Mcp-Session-Id: abc123 
@sessions["abc123"][:stream] = attacker_stream `# Victim's stream is REPLACED (silently disconnected)

Step 4: Data Interception

  • ALL subsequent tool responses/notifications go to the attacker
  • The legitimate user receives no data and has no indication of the hijacking

Technical Details

The vulnerability happens:

Client 1 connects (GET request)

proc do |stream1|  # ← Rack server provides stream1 for client 1
 @sessions[session_id][:stream] = stream1  # Stored
end

Client 2 connects with SAME session ID (Attack!)

proc do |stream2|  # ← Rack provides stream2 for client 2
 @sessions[session_id][:stream] = stream2  # REPLACES stream1!
end

Now when the server sends notifications:

@sessions[session_id][:stream].write(data)  # Goes to stream2 (attacker!)
# stream1 (victim) receives nothing

Comparison: Python SDK Protection

The Python SDK prevents this vulnerability by rejecting duplicate SSE connections:

Refer: https://github.com/modelcontextprotocol/python-sdk/blob/main/src/mcp/server/streamable_http.py#L680-L685

if GET_STREAM_KEY in self._request_streams:  # pragma: no cover
            response = self._create_error_response(
                "Conflict: Only one SSE stream is allowed per session",
                HTTPStatus.CONFLICT,
            )

When a duplicate connection attempt is detected, the Python SDK returns an HTTP 409 Conflict error, protecting the existing connection.

Recommended Mitigations For SDK Maintainers

  • Implement User Binding: All SDKs should bind session IDs to authenticated user identities where possible. Currently only, go-sdk and csharp-sdk do user binding.
  • Ruby SDK: Prevent Duplicate Connections: Implement checks to reject or handle multiple simultaneous connections to the same session
  • Improve Documentation: Provide clear guidance on secure session management implementation for SDK consumers

Steps To Reproduce:

Please find attached two python client files demonstrating the attack

Terminal 1: ruby streamable_http_server.rb

Makes use of https://github.com/modelcontextprotocol/ruby-sdk/blob/main/examples/streamable_http_server.rb This server has a tool call notification_tool which the clients call

Terminal 2:

python3 legitimate_client_ruby_server.py

What happens:

  • The client connects and prints the session ID
  • Press Enter to start the SSE stream
  • Notifications start appearing every 3 seconds as the client makes a tool call

Terminal 3 (while the legitimate client is running):

python3 attacker_client_ruby_server.py <SESSION_ID>

Replace <SESSION_ID> with the ID from Terminal 2.

What happens immediately:

  • Terminal 2 (Legitimate): Stops receiving notifications, shows disconnect message
  • Terminal 3 (Attacker): Starts receiving ALL the tool call responses

Impact

While the absence of user binding may not pose immediate risks if session IDs are not used to store sensitive data or state, the fundamental purpose of session IDs is to maintain stateful connections. If the SDK or its consumers utilize session IDs for sensitive operations without proper user binding controls, this creates a potential security vulnerability. For example: In the case of the Ruby SDK, the attacker was able to hijack the stream and receive all the tool responses belonging to the victim. The tool responses can be sensitive confidential data.

Additional Details

Session Hijacking Protection in MCP Implementations

The MCP specification recommends - "MCP servers SHOULD bind session IDs to user-specific information".

Current Implementation Status Across SDKs

Of the 10 official MCP SDKs, only the following implementations bind session IDs to user-specific information:

  1. csharp-sdk - https://github.com/modelcontextprotocol/csharp-sdk/blob/main/src/ModelContextProtocol.AspNetCore/SseHandler.cs#L93-L97
  2. Go-sdk - https://github.com/modelcontextprotocol/go-sdk/blob/main/mcp/streamable.go#L281C1-L288C2

attacker_client_ruby_server.py legitimate_client_ruby_server.py The remaining SDKs do not implement session-to-user binding. Most implementations only verify that a session ID exists, without validating ownership. Additionally, SDK documentation does not provide clear guidance on implementing secure session management, leaving security responsibilities unclear for SDK consumers.

Severity
8.2 (High)
CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
Show details on source website
To clipboard 

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 0.9.1"
      },
      "package": {
        "ecosystem": "RubyGems",
        "name": "mcp"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.9.2"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-33946"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-384",
      "CWE-639"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-03-27T18:36:45Z",
    "nvd_published_at": "2026-03-27T22:16:21Z",
    "severity": "HIGH"
  },
  "details": "### Summary\n\nThe Ruby SDK\u0027s [streamable_http_transport.rb](https://github.com/modelcontextprotocol/ruby-sdk/blob/main/lib/mcp/server/transports/streamable_http_transport.rb) implementation contains a session hijacking vulnerability. An attacker who obtains a valid session ID can completely hijack the victim\u0027s Server-Sent Events (SSE) stream and intercept all real-time data.\n\n### Details\n**Root Cause**\nThe StreamableHTTPTransport implementation stores only one SSE stream object per session ID and lacks:\n\n- Session-to-user identity binding\n- Ownership validation when establishing SSE connections\n- Protection against multiple simultaneous connections to the same session\n\n### PoC\n\n#### Vulnerable Code\n\n**File**: streamable_http_transport.rb - [L336-L339](https://github.com/modelcontextprotocol/ruby-sdk/blob/main/lib/mcp/server/transports/streamable_http_transport.rb#L336-L339):\n\n```\ndef store_stream_for_session(session_id, stream)\n  @mutex.synchronize do\n    if @sessions[session_id]\n      @sessions[session_id][:stream] = stream  # OVERWRITES existing stream\n    else\n      stream.close\n    end\n  end\nend\n```\n#### Attack Scenario\n**Step 1**: Legitimate Session Establishment\n```\nPOST / (initialize) \u2192 receives session_id: \"abc123\"\nGET / with Mcp-Session-Id: abc123 \u2192 SSE stream connected\n```\nStep 2: Session ID Compromise\n\n- An attacker obtains the session ID through various means (out of scope for this analysis)\n\n**Step 3**: Stream Hijacking\n\n```\nGET / with Mcp-Session-Id: abc123 \n@sessions[\"abc123\"][:stream] = attacker_stream `# Victim\u0027s stream is REPLACED (silently disconnected)\n```\n\n**Step 4**: Data Interception\n\n- ALL subsequent tool responses/notifications go to the attacker\n- The legitimate user receives no data and has no indication of the hijacking\n\n#### Technical Details\n\nThe vulnerability happens:\n\n**Client 1 connects (GET request)**\n\n```\nproc do |stream1|  # \u2190 Rack server provides stream1 for client 1\n @sessions[session_id][:stream] = stream1  # Stored\nend\n```\n\n**Client 2 connects with SAME session ID (Attack!)**\n```\nproc do |stream2|  # \u2190 Rack provides stream2 for client 2\n @sessions[session_id][:stream] = stream2  # REPLACES stream1!\nend\n```\n\n**Now when the server sends notifications:**\n\n```\n@sessions[session_id][:stream].write(data)  # Goes to stream2 (attacker!)\n# stream1 (victim) receives nothing\n```\n\n**Comparison: Python SDK Protection**\n\nThe Python SDK prevents this vulnerability by rejecting duplicate SSE connections:\n\n**Refer**: https://github.com/modelcontextprotocol/python-sdk/blob/main/src/mcp/server/streamable_http.py#L680-L685\n\n```\nif GET_STREAM_KEY in self._request_streams:  # pragma: no cover\n            response = self._create_error_response(\n                \"Conflict: Only one SSE stream is allowed per session\",\n                HTTPStatus.CONFLICT,\n            )\n```\n\nWhen a duplicate connection attempt is detected, the Python SDK returns an HTTP 409 Conflict error, protecting the existing connection.\n\n**Recommended Mitigations**\n**For SDK Maintainers**\n\n- Implement User Binding: All SDKs should bind session IDs to authenticated user identities where possible. Currently only, go-sdk and csharp-sdk do user binding.\n- **Ruby SDK**: Prevent Duplicate Connections: Implement checks to reject or handle multiple simultaneous connections to the same session\n- **Improve Documentation**: Provide clear guidance on secure session management implementation for SDK consumers\n\n### Steps To Reproduce:\n\nPlease find attached two python client files demonstrating the attack\n\n**Terminal 1:**\n`ruby streamable_http_server.rb`\n\nMakes use of https://github.com/modelcontextprotocol/ruby-sdk/blob/main/examples/streamable_http_server.rb\nThis server has a tool call notification_tool which the clients call\n\n**Terminal 2:**\n\n`python3 legitimate_client_ruby_server.py`\n\n**What happens:**\n\n- The client connects and prints the session ID\n- Press Enter to start the SSE stream\n- Notifications start appearing every 3 seconds as the client makes a tool call\n\n**Terminal 3 (while the legitimate client is running):**\n\n`python3 attacker_client_ruby_server.py \u003cSESSION_ID\u003e`\n\nReplace `\u003cSESSION_ID\u003e` with the ID from Terminal 2.\n\n**What happens immediately:**\n\n- Terminal 2 (Legitimate): Stops receiving notifications, shows disconnect message\n- Terminal 3 (Attacker): Starts receiving ALL the tool call responses\n\n### Impact\nWhile the absence of user binding may not pose immediate risks if session IDs are not used to store sensitive data or state, the fundamental purpose of session IDs is to maintain stateful connections. If the SDK or its consumers utilize session IDs for sensitive operations without proper user binding controls, this creates a potential security vulnerability. For example: In the case of the Ruby SDK, the attacker was able to hijack the stream and receive all the tool responses belonging to the victim. The tool responses can be sensitive confidential data.\n\n### Additional Details\n#### Session Hijacking Protection in MCP Implementations\nThe MCP specification recommends - \"MCP servers SHOULD bind session IDs to user-specific information\".\n\n#### Current Implementation Status Across SDKs\n\nOf the 10 official MCP SDKs, only the following implementations bind session IDs to user-specific information:\n\n1. csharp-sdk - https://github.com/modelcontextprotocol/csharp-sdk/blob/main/src/ModelContextProtocol.AspNetCore/SseHandler.cs#L93-L97\n2. Go-sdk - https://github.com/modelcontextprotocol/go-sdk/blob/main/mcp/streamable.go#L281C1-L288C2\n\n[attacker_client_ruby_server.py](https://github.com/user-attachments/files/25408485/attacker_client_ruby_server.py)\n[legitimate_client_ruby_server.py](https://github.com/user-attachments/files/25408486/legitimate_client_ruby_server.py)\nThe remaining SDKs do not implement session-to-user binding. Most implementations only verify that a session ID exists, without validating ownership. Additionally, SDK documentation does not provide clear guidance on implementing secure session management, leaving security responsibilities unclear for SDK consumers.",
  "id": "GHSA-qvqr-5cv7-wh35",
  "modified": "2026-03-30T20:10:39Z",
  "published": "2026-03-27T18:36:45Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/modelcontextprotocol/ruby-sdk/security/advisories/GHSA-qvqr-5cv7-wh35"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-33946"
    },
    {
      "type": "WEB",
      "url": "https://github.com/modelcontextprotocol/ruby-sdk/commit/db40143402d65b4fb6923cec42d2d72cb89b3874"
    },
    {
      "type": "WEB",
      "url": "https://hackerone.com/reports/3556146"
    },
    {
      "type": "WEB",
      "url": "https://github.com/modelcontextprotocol/csharp-sdk/blob/main/src/ModelContextProtocol.AspNetCore/SseHandler.cs#L93-L97"
    },
    {
      "type": "WEB",
      "url": "https://github.com/modelcontextprotocol/go-sdk/blob/main/mcp/streamable.go#L281C1-L288C2"
    },
    {
      "type": "WEB",
      "url": "https://github.com/modelcontextprotocol/python-sdk/blob/main/src/mcp/server/streamable_http.py#L680-L685"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/modelcontextprotocol/ruby-sdk"
    },
    {
      "type": "WEB",
      "url": "https://github.com/modelcontextprotocol/ruby-sdk/blob/main/examples/streamable_http_server.rb"
    },
    {
      "type": "WEB",
      "url": "https://github.com/modelcontextprotocol/ruby-sdk/releases/tag/v0.9.2"
    },
    {
      "type": "WEB",
      "url": "https://github.com/rubysec/ruby-advisory-db/blob/master/gems/mcp/CVE-2026-33946.yml"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N",
      "type": "CVSS_V4"
    }
  ],
  "summary": "MCP Ruby SDK: Insufficient Session Binding Allows SSE Stream Hijacking via Session ID Replay"
}

GHSA-QW5P-3FQ9-8GGV

Vulnerability from github – Published: 2022-05-24 19:14 – Updated: 2022-05-24 19:14
VLAI
Details

IBM Security Guardium 10.6 and 11.3 could allow a remote authenticated attacker to obtain sensitive information or modify user details caused by an insecure direct object vulnerability (IDOR). IBM X-Force ID: 202865.

Severity
5.4 (Medium)
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N
Show details on source website
To clipboard 

{
  "affected": [],
  "aliases": [
    "CVE-2021-29773"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-200",
      "CWE-639"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2021-09-15T18:15:00Z",
    "severity": "MODERATE"
  },
  "details": "IBM Security Guardium 10.6 and 11.3 could allow a remote authenticated attacker to obtain sensitive information or modify user details caused by an insecure direct object vulnerability (IDOR). IBM X-Force ID: 202865.",
  "id": "GHSA-qw5p-3fq9-8ggv",
  "modified": "2022-05-24T19:14:36Z",
  "published": "2022-05-24T19:14:36Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2021-29773"
    },
    {
      "type": "WEB",
      "url": "https://exchange.xforce.ibmcloud.com/vulnerabilities/202865"
    },
    {
      "type": "WEB",
      "url": "https://www.ibm.com/support/pages/node/6488943"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-QW84-4PC7-FXVW

Vulnerability from github – Published: 2026-04-14 09:30 – Updated: 2026-04-14 09:30
VLAI
Details

A vulnerability has been identified in SINEC NMS (All versions < V4.0 SP3). Affected products do not properly validate user authorization when processing password reset requests. This could allow an authenticated remote attacker to bypass authorization checks, leading to the ability to reset the password of any arbitrary user account.

Severity
8.8 (High)
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
8.7 (High)
CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
Show details on source website
To clipboard 

{
  "affected": [],
  "aliases": [
    "CVE-2026-25654"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-639"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-04-14T09:16:35Z",
    "severity": "HIGH"
  },
  "details": "A vulnerability has been identified in SINEC NMS (All versions \u003c V4.0 SP3). Affected products do not properly validate user authorization when processing password reset requests. This could allow an authenticated remote attacker to bypass authorization checks, leading to the ability to reset the password of any arbitrary user account.",
  "id": "GHSA-qw84-4pc7-fxvw",
  "modified": "2026-04-14T09:30:44Z",
  "published": "2026-04-14T09:30:44Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-25654"
    },
    {
      "type": "WEB",
      "url": "https://cert-portal.siemens.com/productcert/html/ssa-605717.html"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    },
    {
      "score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
      "type": "CVSS_V4"
    }
  ]
}

GHSA-QWXF-2M7M-2M3X

Vulnerability from github – Published: 2026-06-17 18:07 – Updated: 2026-07-20 21:10
VLAI
Summary
Daytona: Cross-tenant data leak in notification WebSocket gateway via unverified organizationId join
Details

Summary

A cross-tenant authorization flaw in Daytona's notification WebSocket gateway allowed any authenticated user to subscribe to another organization's realtime notification channel and passively receive that organization's events.

Impact

The notification gateway's JWT handshake joined a client-supplied organization identifier to the corresponding notification room without verifying that the authenticated user was a member of that organization. As a result, an authenticated user could receive another organization's realtime sandbox, snapshot, volume, and runner events, including data carried in those events. This is a cross-tenant confidentiality break. It required a valid account and knowledge of the target organization id (a non-secret UUID); no elevated privileges were needed. The API-key authentication path was not affected.

The affected component is the Daytona API service (the apps/api NestJS application). It is distributed through Daytona's repository releases and container images for self-hosted deployments; it is not published as a Go or npm package, so the advisory will not surface through go get or npm dependency tooling.

Affected Versions

= 0.101.0, <= 0.184.0

Patched Versions

0.185.0

Credit

@vnth4nhnt from CyStack

Severity
6.5 (Medium)
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N
Show details on source website
To clipboard 

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 0.184.0"
      },
      "package": {
        "ecosystem": "Go",
        "name": "github.com/daytonaio/daytona"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.185.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-54324"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-639",
      "CWE-863"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-17T18:07:30Z",
    "nvd_published_at": "2026-06-23T18:18:09Z",
    "severity": "MODERATE"
  },
  "details": "### Summary\nA cross-tenant authorization flaw in Daytona\u0027s notification WebSocket gateway allowed any authenticated user to subscribe to another organization\u0027s realtime notification channel and passively receive that organization\u0027s events.\n\n### Impact\nThe notification gateway\u0027s JWT handshake joined a client-supplied organization identifier to the corresponding notification room without verifying that the authenticated user was a member of that organization. As a result, an authenticated user could receive another organization\u0027s realtime sandbox, snapshot, volume, and runner events, including data carried in those events. This is a cross-tenant confidentiality break. It required a valid account and knowledge of the target organization id (a non-secret UUID); no elevated privileges were needed. The API-key authentication path was not affected.\n\nThe affected component is the Daytona API service (the `apps/api` NestJS application). It is distributed through Daytona\u0027s repository releases and container images for self-hosted deployments; it is not published as a Go or npm package, so the advisory will not surface through `go get` or npm dependency tooling.\n\n### Affected Versions\n\u003e= 0.101.0, \u003c= 0.184.0\n\n### Patched Versions\n0.185.0\n\n### Credit\n@vnth4nhnt from CyStack",
  "id": "GHSA-qwxf-2m7m-2m3x",
  "modified": "2026-07-20T21:10:57Z",
  "published": "2026-06-17T18:07:30Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/daytonaio/daytona/security/advisories/GHSA-qwxf-2m7m-2m3x"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-54324"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/daytonaio/daytona"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Daytona: Cross-tenant data leak in notification WebSocket gateway via unverified organizationId join"
}

GHSA-QX86-G93J-M25R

Vulnerability from github – Published: 2026-04-23 15:38 – Updated: 2026-04-23 15:38
VLAI
Details

An API design flaw in WebKitGTK and WPE WebKit allows untrusted web content to unexpectedly perform IP connections, DNS lookups, and HTTP requests. Applications expect to use the WebPage::send-request signal handler to approve or reject all network requests. However, certain types of HTTP requests bypass this signal handler.

Severity
4.7 (Medium)
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:N/A:N
Show details on source website
To clipboard 

{
  "affected": [],
  "aliases": [
    "CVE-2025-66286"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-639"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-04-23T13:16:11Z",
    "severity": "MODERATE"
  },
  "details": "An API design flaw in WebKitGTK and WPE WebKit allows untrusted web content to unexpectedly perform IP connections, DNS lookups, and HTTP requests. Applications expect to use the\nWebPage::send-request signal handler to approve or reject all network requests. However, certain types of HTTP requests bypass this signal handler.",
  "id": "GHSA-qx86-g93j-m25r",
  "modified": "2026-04-23T15:38:56Z",
  "published": "2026-04-23T15:38:56Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-66286"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/security/cve/CVE-2025-66286"
    },
    {
      "type": "WEB",
      "url": "https://bugs.webkit.org/show_bug.cgi?id=259787"
    },
    {
      "type": "WEB",
      "url": "https://bugzilla.redhat.com/show_bug.cgi?id=2424652"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-QXM8-6XVG-W483

Vulnerability from github – Published: 2025-07-18 15:31 – Updated: 2026-06-01 15:30
VLAI
Details

Authorization Bypass Through User-Controlled Key vulnerability in Vidco Software VOC TESTER allows Forceful Browsing.This issue affects VOC TESTER: before 12.41.0.

Severity
5.5 (Medium)
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N
Show details on source website
To clipboard 

{
  "affected": [],
  "aliases": [
    "CVE-2024-13175"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-639"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2025-07-18T14:15:23Z",
    "severity": "MODERATE"
  },
  "details": "Authorization Bypass Through User-Controlled Key vulnerability in Vidco Software VOC TESTER allows Forceful Browsing.This issue affects VOC TESTER: before 12.41.0.",
  "id": "GHSA-qxm8-6xvg-w483",
  "modified": "2026-06-01T15:30:32Z",
  "published": "2025-07-18T15:31:56Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2024-13175"
    },
    {
      "type": "WEB",
      "url": "https://siberguvenlik.gov.tr/guvenlik-bildirimleri/detay/tr-25-0159"
    },
    {
      "type": "WEB",
      "url": "https://www.usom.gov.tr/bildirim/tr-25-0159"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-QXQC-G59M-CQX4

Vulnerability from github – Published: 2026-02-03 09:30 – Updated: 2026-02-03 09:30
VLAI
Details

The Tutor LMS – eLearning and online course solution plugin for WordPress is vulnerable to Insecure Direct Object References (IDOR) in all versions up to, and including, 3.9.5. This is due to missing object-level authorization checks in the course_list_bulk_action(), bulk_delete_course(), and update_course_status() functions. This makes it possible for authenticated attackers, with Tutor Instructor-level access and above, to modify or delete arbitrary courses they do not own by manipulating course IDs in bulk action requests.

Severity
8.1 (High)
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H
Show details on source website
To clipboard 

{
  "affected": [],
  "aliases": [
    "CVE-2026-1375"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-639"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-02-03T08:16:14Z",
    "severity": "HIGH"
  },
  "details": "The Tutor LMS \u2013 eLearning and online course solution plugin for WordPress is vulnerable to Insecure Direct Object References (IDOR) in all versions up to, and including, 3.9.5. This is due to missing object-level authorization checks in the `course_list_bulk_action()`, `bulk_delete_course()`, and `update_course_status()` functions. This makes it possible for authenticated attackers, with Tutor Instructor-level access and above, to modify or delete arbitrary courses they do not own by manipulating course IDs in bulk action requests.",
  "id": "GHSA-qxqc-g59m-cqx4",
  "modified": "2026-02-03T09:30:28Z",
  "published": "2026-02-03T09:30:28Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-1375"
    },
    {
      "type": "WEB",
      "url": "https://plugins.trac.wordpress.org/browser/tutor/tags/3.9.5/classes/Course_List.php#L289"
    },
    {
      "type": "WEB",
      "url": "https://plugins.trac.wordpress.org/browser/tutor/tags/3.9.5/classes/Course_List.php#L437"
    },
    {
      "type": "WEB",
      "url": "https://plugins.trac.wordpress.org/browser/tutor/tags/3.9.5/classes/Course_List.php#L463"
    },
    {
      "type": "WEB",
      "url": "https://plugins.trac.wordpress.org/changeset/3448615/tutor/trunk/classes/Course_List.php?contextall=1\u0026old=3339576\u0026old_path=%2Ftutor%2Ftrunk%2Fclasses%2FCourse_List.php"
    },
    {
      "type": "WEB",
      "url": "https://www.wordfence.com/threat-intel/vulnerabilities/id/4e95b32b-c050-41eb-8fce-461257420eb6?source=cve"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-QXVH-QPPG-J3PV

Vulnerability from github – Published: 2026-06-23 18:31 – Updated: 2026-06-23 18:31
VLAI
Details

Pega Platform versions 8.3.0 through Infinity 25.1.2 are affected by an authorization weakness that may allow authenticated users to access certain additional data via crafted URLs.

Severity
7.1 (High)
CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:L/SC:N/SI:N/SA:N
Show details on source website
To clipboard 

{
  "affected": [],
  "aliases": [
    "CVE-2025-62180"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-639"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-06-23T16:16:58Z",
    "severity": "HIGH"
  },
  "details": "Pega Platform versions 8.3.0 through Infinity 25.1.2 are affected by an authorization weakness that may allow authenticated users to access certain additional data via crafted URLs.",
  "id": "GHSA-qxvh-qppg-j3pv",
  "modified": "2026-06-23T18:31:37Z",
  "published": "2026-06-23T18:31:37Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-62180"
    },
    {
      "type": "WEB",
      "url": "https://support.pega.com/support-doc/pega-security-advisory-h26-vulnerability-remediation-note"
    },
    {
      "type": "WEB",
      "url": "https://support.pega.com/support-doc/pega-security-advisory-i25-vulnerability-remediation-note"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
      "type": "CVSS_V4"
    }
  ]
}

GHSA-QXVM-PCFM-QC39

Vulnerability from github – Published: 2026-06-16 21:30 – Updated: 2026-07-20 21:15
VLAI
Summary
Daytona: Cross-org IDOR in organization role update/delete — any org owner can rewrite or destroy another org's roles
Details

Summary

Daytona's organization role update and delete endpoints authorized the caller as an owner of the organization named in the request path, but resolved and mutated the target role by its identifier alone, without verifying the role belonged to that organization. An authenticated user who owns any organization (organizations are self-service) could therefore modify the permissions of, or delete, a role belonging to a different organization, given that role's identifier.

Impact

This is a cross-tenant broken access control (IDOR) issue affecting multi-tenant deployments, including the managed Daytona platform. Using a target role's identifier, an attacker with owner rights over their own organization could:

  • Overwrite the target role's name and permission set, escalating or stripping privileges for every member and API key in the victim organization that holds that role.
  • Delete the target role, removing the associated permissions from its holders.
  • Observe the victim role's current permission set returned in the update response (limited information disclosure).

Exploitation requires knowledge of the target role's identifier, which is not enumerable across organizations and is not exposed to non-members through the API.

Affected versions

All versions up to and including 0.184.0.

Patches

Fixed in 0.185.0. The role update, delete, and role-assignment lookups are now scoped to the caller's organization, so a role belonging to another organization resolves to "not found" before any read or mutation. The managed Daytona platform was updated on release of 0.185.0.

Workarounds

None. Upgrade to 0.185.0. Single-organization self-hosted deployments are not exploitable, as the issue requires a second organization to target.

Credit

Reported by @vnth4nhnt.

Severity
7.7 (High)
CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:H/A:L
Show details on source website
To clipboard 

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 0.184.0"
      },
      "package": {
        "ecosystem": "Go",
        "name": "github.com/daytonaio/daytona"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.185.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-54322"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-639",
      "CWE-862"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-16T21:30:08Z",
    "nvd_published_at": "2026-06-23T19:17:08Z",
    "severity": "HIGH"
  },
  "details": "### Summary\nDaytona\u0027s organization role update and delete endpoints authorized the caller as an owner of the organization named in the request path, but resolved and mutated the target role by its identifier alone, without verifying the role belonged to that organization. An authenticated user who owns any organization (organizations are self-service) could therefore modify the permissions of, or delete, a role belonging to a different organization, given that role\u0027s identifier.\n\n### Impact\nThis is a cross-tenant broken access control (IDOR) issue affecting multi-tenant deployments, including the managed Daytona platform. Using a target role\u0027s identifier, an attacker with owner rights over their own organization could:\n\n- Overwrite the target role\u0027s name and permission set, escalating or stripping privileges for every member and API key in the victim organization that holds that role.\n- Delete the target role, removing the associated permissions from its holders.\n- Observe the victim role\u0027s current permission set returned in the update response (limited information disclosure).\n\nExploitation requires knowledge of the target role\u0027s identifier, which is not enumerable across organizations and is not exposed to non-members through the API.\n\n### Affected versions\nAll versions up to and including 0.184.0.\n\n### Patches\nFixed in 0.185.0. The role update, delete, and role-assignment lookups are now scoped to the caller\u0027s organization, so a role belonging to another organization resolves to \"not found\" before any read or mutation. The managed Daytona platform was updated on release of 0.185.0.\n\n### Workarounds\nNone. Upgrade to 0.185.0. Single-organization self-hosted deployments are not exploitable, as the issue requires a second organization to target.\n\n### Credit\nReported by @vnth4nhnt.",
  "id": "GHSA-qxvm-pcfm-qc39",
  "modified": "2026-07-20T21:15:04Z",
  "published": "2026-06-16T21:30:08Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/daytonaio/daytona/security/advisories/GHSA-qxvm-pcfm-qc39"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-54322"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/daytonaio/daytona"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:H/A:L",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Daytona: Cross-org IDOR in organization role update/delete \u2014 any org owner can rewrite or destroy another org\u0027s roles"
}

GHSA-QXVM-R42F-5P8J

Vulnerability from github – Published: 2026-05-15 18:17 – Updated: 2026-05-15 18:17
VLAI
Summary
AVideo's Meet plugin: `uploadRecordedVideo.json.php` derives `users_id` from the uploaded filename and calls passwordless `User->login()`, allowing any caller with the Meet shared secret to obtain a session as arbitrary users including admin
Details

Summary

Type: Authorization-bypass via user-controlled identifier. The Meet plugin's recorded-video upload endpoint (plugin/Meet/uploadRecordedVideo.json.php) authenticates the caller using a single shared Authorization: Bearer <secret> against $objM->secret. Once that check passes, the endpoint reads the target user identifier from the uploaded file's name field, instantiates a User object with that ID, and calls $userObject->login(true, true) — the no-password / encoded-password login path — committing a session for that user and emitting Set-Cookie headers to the caller. There is no check that the caller actually owns the requested users_id. File: plugin/Meet/uploadRecordedVideo.json.php, lines 56-65; secondary in objects/user.php User::login() (no-password branch at lines 1276-1310). Root cause: the upload handler's identity model is "service-to-service" (a Meet/Jitsi recorder posts a finished recording back to AVideo with the shared secret) but the users_id to credit the upload to is parsed from the FILENAME the same caller controls — $users_id = explode('-', $_FILES['upl']['name'])[0];. There is no signed claim, no separate proof-of-identity, no allowlist. The subsequent $userObject->login(true, true) call invokes the no-password login path which sets $_SESSION['user'], calls setUserCookie(...), and _session_regenerate_id() — exactly the operations a normal login performs. The response carries the new PHPSESSID back to the caller, who can then reuse it on every subsequent request to act as the targeted user. The Meet shared secret is md5($global['systemRootPath'] . $global['salt'] . "meet") (Meet.php:73), so any attacker who can read videos/configuration.php (e.g., via a path-traversal CVE such as GHSA-83xq-8jxj-4rxm or GHSA-4wmm-6qxj-fpj4 that the project has already addressed in this surface area) can compute the Meet secret deterministically and pivot to full account takeover.

Affected Code

File: plugin/Meet/uploadRecordedVideo.json.php, lines 33-73.

if (empty($token)) {
    forbiddenPage('Token not found');
}

$objM = AVideoPlugin::getObjectDataIfEnabled("Meet");
if (empty($objM)) {
    forbiddenPage('Plugin disabled');
}

if ($objM->secret != $token) {                              // <-- shared-secret auth, no per-user proof
    forbiddenPage('Token does not match');
}

if (empty($_FILES['upl'])) {
    forbiddenPage('videoFile not found');
}

$users_id = explode('-', $_FILES['upl']['name'])[0];        // <-- BUG: target users_id parsed from attacker-controlled filename

$userObject = new User($users_id);
$userObject->login(true, true);                             // <-- BUG: passwordless login as the chosen user; sets $_SESSION + Set-Cookie
$tmpFile = getTmpDir() . uniqid();

if (move_uploaded_file($_FILES['upl']['tmp_name'], $tmpFile)) {
    $_FILES['upl']['tmp_name'] = $tmpFile;
    require $global['systemRootPath'] . 'objects/aVideoQueueEncoder.json.php';
}

File: objects/user.php, lines 1249-1329 (User::login() no-password branch).

public function login($noPass = false, $encodedPass = false, $ignoreEmailVerification = false)
{
    // ...
    if ($noPass) {
        $user = $this->find($this->user, false, true);      // <-- no password check
    }
    // ...
    } elseif ($user) {
        $_SESSION['user'] = $user;                          // <-- session set for the impersonated user
        $this->setLastLogin($_SESSION['user']['id']);
        // ...
        self::setUserCookie($rememberme, $user['id'], $user['user'], $passhash, $expires);
        AVideoPlugin::onUserSignIn($_SESSION['user']['id']);
        $_SESSION['loginAttempts'] = 0;
        _session_regenerate_id();                           // <-- new SID committed in Set-Cookie response
        _session_write_close();
        return self::USER_LOGGED;
    }
}

Why it's wrong: the endpoint conflates two distinct authentication concerns. The shared-secret check answers "is this request coming from a trusted Meet recorder?" but the filename parse answers "which user does this recording belong to?" — and the second answer is taken from the same untrusted caller. Once User->login(true, true) runs, the server has no way to distinguish a legitimate Meet integration from an attacker who happens to know the same secret. The decision to expose this as a session (cookie + _session_regenerate_id) rather than as a one-shot in-process credit makes the impact larger than it needs to be: even if the Meet integration only needed to credit the recording to a user, the implementation gives the caller a fully-authenticated session as that user.

Exploit Chain

  1. Attacker obtains the Meet shared secret. Two plausible paths:
  2. Path A (computational): the secret is md5($global['systemRootPath'] . $global['salt'] . "meet") (plugin/Meet/Meet.php:73). Both inputs sit in videos/configuration.php. AVideo's history of LFI/path-traversal CVEs in this surface (e.g., the import.json.php and listFiles.json.php advisories already accepted on this program) means the salt is a realistic disclosure target.
  3. Path B (timing oracle): plugin/Meet/checkToken.json.php line 26 does if ($objM->secret === $_GET['secret']) with no constant-time comparison and a clear yes/no response body. PHP's === for strings short-circuits on first byte mismatch, so an attacker on the same network segment can recover the 32-hex secret byte-by-byte over the network with timing analysis. Slower than path A but doesn't depend on a separate vulnerability.
  4. Attacker prepares an HTTP POST to /plugin/Meet/uploadRecordedVideo.json.php:
  5. Authorization: Bearer <Meet secret>
  6. Multipart body with one file field named upl. The filename is set to 1-anything.mp4 (where 1 is the users_id of the admin or any target user — the format is <users_id>-<arbitrary>). The file body itself can be anything that survives the surrounding aVideoQueueEncoder pipeline (an empty file is enough to reach the login call before the encoder rejects).
  7. Server flow:
  8. Line 33: token present, ok.
  9. Line 46: $objM->secret != $token → false (matches), passes.
  10. Line 51: $_FILES['upl'] present, ok.
  11. Line 56: $users_id = explode('-', '1-anything.mp4')[0] → '1'.
  12. Line 59-60: $userObject = new User(1); $userObject->login(true, true); — passwordless login as user 1 (admin). $_SESSION['user'] is set, setUserCookie runs, _session_regenerate_id issues a new session ID, and the response carries Set-Cookie: PHPSESSID=<new-sid>; ....
  13. Subsequent code runs the encoder pipeline as admin — but the attacker's primary goal was already achieved when the session was established.
  14. Attacker captures the Set-Cookie: PHPSESSID=... header from the response and uses that cookie on all subsequent requests. Server treats them as user 1 (admin) — full UI access, all admin endpoints, all video management, plugin configuration, user impersonation, etc.
  15. Final state: admin account takeover. The original Meet recorder's flow (legitimate uploads with users_id = the user who scheduled the meeting) is indistinguishable on the wire from the attack flow (users_id = whoever the attacker wants to be).

Security Impact

Severity: sec-high. End state is full account takeover of any user (including admin), reachable from a single HTTP POST once the secret is known. The shared-secret precondition raises AC to High but does not eliminate it as a credible threat — the secret is computable from any leak of videos/configuration.php, and AVideo's CVE history in that surface area is non-trivial. Attacker capability: session hijack as any users_id the attacker cares to name. The attacker chooses the target by setting the filename's leading digits before the first -. No bound on which user IDs are reachable; admin (1 on a default install) is the obvious target. Once the session is captured, the attacker has full admin UI/API access for the session lifetime (hours-to-days depending on rememberme flag). Preconditions: Meet plugin enabled (default-off but commonly enabled by deployments using AVideo for video-conferencing recording). Knowledge of the Meet shared secret (computable from the salt; obtainable via timing attack on checkToken.json.php). Differential: source-inspection-verified end-to-end. The two relevant code blocks are quoted verbatim in §Affected Code; both lines are reachable on every successful POST to the endpoint. The patched build (with the suggested fix below) either rejects the upload as 'cannot derive identity from filename' or constrains the users_id to one bound by an additional signed claim from the Meet recorder.

Suggested Fix

Three changes, in order of importance:

--- a/plugin/Meet/uploadRecordedVideo.json.php
+++ b/plugin/Meet/uploadRecordedVideo.json.php
@@ -53,17 +53,28 @@ if (empty($_FILES['upl'])) {
     forbiddenPage('videoFile not found');
 }

-$users_id = explode('-', $_FILES['upl']['name'])[0];
+// The users_id MUST come from a signed claim (e.g., a JWT issued by AVideo
+// when the meeting was scheduled), not from a filename the caller controls.
+// Verify a recording-upload token here that was minted at meeting-create
+// time and bound to (meet_schedule_id, users_id) with an HMAC.
+$claim = MeetUploadClaim::verifyFromHeaders($headers);
+if (!$claim) {
+    forbiddenPage('Missing or invalid recording upload claim');
+}
+$users_id = (int) $claim->users_id;
+if (!$users_id || !User::idExists($users_id)) {
+    forbiddenPage('Recording upload claim references unknown user');
+}

-$userObject = new User($users_id);
-$userObject->login(true, true);
+// Credit the upload to $users_id WITHOUT establishing a session. The encoder
+// pipeline can be parameterised to record ownership directly; there is no
+// reason for a service-to-service upload endpoint to mint a user session.
+$queueOwnerUsersId = $users_id;
 $tmpFile = getTmpDir() . uniqid();

 if (move_uploaded_file($_FILES['upl']['tmp_name'], $tmpFile)) {
     $_FILES['upl']['tmp_name'] = $tmpFile;
-    require $global['systemRootPath'] . 'objects/aVideoQueueEncoder.json.php';
+    aVideoQueueEncoder::encodeOnBehalfOf($queueOwnerUsersId, $_FILES['upl']);
 }

Additionally:

  1. Use hash_equals for the secret comparison in both this endpoint and checkToken.json.php (if (!hash_equals($objM->secret, $token))). The current ==/=== is vulnerable to byte-by-byte timing analysis.
  2. Remove checkToken.json.php entirely, or at least gate it behind User::isAdmin(). A network-reachable endpoint that confirms whether a guess matches the server-side secret is exactly the wrong shape for a high-value secret like this one.

Optional defense-in-depth (separate change): rotate the Meet secret to use a random 256-bit value (not derived from salt), so a videos/configuration.php disclosure does not also yield the Meet secret. Store the random secret as a per-deployment row in the Meet plugin's configuration table, generated at first-run.

Add a regression test: call uploadRecordedVideo.json.php with the correct secret but a filename of 1-x.mp4; assert the response does NOT include a Set-Cookie: PHPSESSID= header.

Severity
8.1 (High)
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H
Show details on source website
To clipboard 

{
  "affected": [
    {
      "package": {
        "ecosystem": "Packagist",
        "name": "WWBN/AVideo"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "last_affected": "29.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [],
  "database_specific": {
    "cwe_ids": [
      "CWE-1390",
      "CWE-287",
      "CWE-639"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-05-15T18:17:19Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "## Summary\n\n**Type:** Authorization-bypass via user-controlled identifier. The Meet plugin\u0027s recorded-video upload endpoint (`plugin/Meet/uploadRecordedVideo.json.php`) authenticates the caller using a single shared `Authorization: Bearer \u003csecret\u003e` against `$objM-\u003esecret`. Once that check passes, the endpoint reads the *target user identifier* from the uploaded file\u0027s `name` field, instantiates a `User` object with that ID, and calls `$userObject-\u003elogin(true, true)` \u2014 the no-password / encoded-password login path \u2014 committing a session for that user and emitting `Set-Cookie` headers to the caller. There is no check that the caller actually owns the requested `users_id`.\n**File:** `plugin/Meet/uploadRecordedVideo.json.php`, lines 56-65; secondary in `objects/user.php` `User::login()` (no-password branch at lines 1276-1310).\n**Root cause:** the upload handler\u0027s identity model is \"service-to-service\" (a Meet/Jitsi recorder posts a finished recording back to AVideo with the shared secret) but the `users_id` to credit the upload to is parsed from the FILENAME the same caller controls \u2014 `$users_id = explode(\u0027-\u0027, $_FILES[\u0027upl\u0027][\u0027name\u0027])[0];`. There is no signed claim, no separate proof-of-identity, no allowlist. The subsequent `$userObject-\u003elogin(true, true)` call invokes the no-password login path which sets `$_SESSION[\u0027user\u0027]`, calls `setUserCookie(...)`, and `_session_regenerate_id()` \u2014 exactly the operations a normal login performs. The response carries the new `PHPSESSID` back to the caller, who can then reuse it on every subsequent request to act as the targeted user. The Meet shared secret is `md5($global[\u0027systemRootPath\u0027] . $global[\u0027salt\u0027] . \"meet\")` (`Meet.php:73`), so any attacker who can read `videos/configuration.php` (e.g., via a path-traversal CVE such as `GHSA-83xq-8jxj-4rxm` or `GHSA-4wmm-6qxj-fpj4` that the project has already addressed in this surface area) can compute the Meet secret deterministically and pivot to full account takeover.\n\n## Affected Code\n\n**File:** `plugin/Meet/uploadRecordedVideo.json.php`, lines 33-73.\n\n```php\nif (empty($token)) {\n    forbiddenPage(\u0027Token not found\u0027);\n}\n\n$objM = AVideoPlugin::getObjectDataIfEnabled(\"Meet\");\nif (empty($objM)) {\n    forbiddenPage(\u0027Plugin disabled\u0027);\n}\n\nif ($objM-\u003esecret != $token) {                              // \u003c-- shared-secret auth, no per-user proof\n    forbiddenPage(\u0027Token does not match\u0027);\n}\n\nif (empty($_FILES[\u0027upl\u0027])) {\n    forbiddenPage(\u0027videoFile not found\u0027);\n}\n\n$users_id = explode(\u0027-\u0027, $_FILES[\u0027upl\u0027][\u0027name\u0027])[0];        // \u003c-- BUG: target users_id parsed from attacker-controlled filename\n\n$userObject = new User($users_id);\n$userObject-\u003elogin(true, true);                             // \u003c-- BUG: passwordless login as the chosen user; sets $_SESSION + Set-Cookie\n$tmpFile = getTmpDir() . uniqid();\n\nif (move_uploaded_file($_FILES[\u0027upl\u0027][\u0027tmp_name\u0027], $tmpFile)) {\n    $_FILES[\u0027upl\u0027][\u0027tmp_name\u0027] = $tmpFile;\n    require $global[\u0027systemRootPath\u0027] . \u0027objects/aVideoQueueEncoder.json.php\u0027;\n}\n```\n\n**File:** `objects/user.php`, lines 1249-1329 (`User::login()` no-password branch).\n\n```php\npublic function login($noPass = false, $encodedPass = false, $ignoreEmailVerification = false)\n{\n    // ...\n    if ($noPass) {\n        $user = $this-\u003efind($this-\u003euser, false, true);      // \u003c-- no password check\n    }\n    // ...\n    } elseif ($user) {\n        $_SESSION[\u0027user\u0027] = $user;                          // \u003c-- session set for the impersonated user\n        $this-\u003esetLastLogin($_SESSION[\u0027user\u0027][\u0027id\u0027]);\n        // ...\n        self::setUserCookie($rememberme, $user[\u0027id\u0027], $user[\u0027user\u0027], $passhash, $expires);\n        AVideoPlugin::onUserSignIn($_SESSION[\u0027user\u0027][\u0027id\u0027]);\n        $_SESSION[\u0027loginAttempts\u0027] = 0;\n        _session_regenerate_id();                           // \u003c-- new SID committed in Set-Cookie response\n        _session_write_close();\n        return self::USER_LOGGED;\n    }\n}\n```\n\n**Why it\u0027s wrong:** the endpoint conflates two distinct authentication concerns. The shared-secret check answers \"is this request coming from a trusted Meet recorder?\" but the filename parse answers \"which user does this recording belong to?\" \u2014 and the second answer is taken from the same untrusted caller. Once `User-\u003elogin(true, true)` runs, the server has no way to distinguish a legitimate Meet integration from an attacker who happens to know the same secret. The decision to expose this as a session (cookie + `_session_regenerate_id`) rather than as a one-shot in-process credit makes the impact larger than it needs to be: even if the Meet integration only needed to *credit* the recording to a user, the implementation gives the caller a fully-authenticated session as that user.\n\n## Exploit Chain\n\n1. Attacker obtains the Meet shared secret. Two plausible paths:\n   - **Path A** (computational): the secret is `md5($global[\u0027systemRootPath\u0027] . $global[\u0027salt\u0027] . \"meet\")` (`plugin/Meet/Meet.php:73`). Both inputs sit in `videos/configuration.php`. AVideo\u0027s history of LFI/path-traversal CVEs in this surface (e.g., the `import.json.php` and `listFiles.json.php` advisories already accepted on this program) means the salt is a realistic disclosure target.\n   - **Path B** (timing oracle): `plugin/Meet/checkToken.json.php` line 26 does `if ($objM-\u003esecret === $_GET[\u0027secret\u0027])` with no constant-time comparison and a clear yes/no response body. PHP\u0027s `===` for strings short-circuits on first byte mismatch, so an attacker on the same network segment can recover the 32-hex secret byte-by-byte over the network with timing analysis. Slower than path A but doesn\u0027t depend on a separate vulnerability.\n2. Attacker prepares an HTTP POST to `/plugin/Meet/uploadRecordedVideo.json.php`:\n   - `Authorization: Bearer \u003cMeet secret\u003e`\n   - Multipart body with one file field named `upl`. The filename is set to `1-anything.mp4` (where `1` is the `users_id` of the admin or any target user \u2014 the format is `\u003cusers_id\u003e-\u003carbitrary\u003e`). The file body itself can be anything that survives the surrounding aVideoQueueEncoder pipeline (an empty file is enough to reach the login call before the encoder rejects).\n3. Server flow:\n   - Line 33: token present, ok.\n   - Line 46: `$objM-\u003esecret != $token` \u2192 false (matches), passes.\n   - Line 51: `$_FILES[\u0027upl\u0027]` present, ok.\n   - Line 56: `$users_id = explode(\u0027-\u0027, \u00271-anything.mp4\u0027)[0]` \u2192 `\u00271\u0027`.\n   - Line 59-60: `$userObject = new User(1); $userObject-\u003elogin(true, true);` \u2014 passwordless login as user 1 (admin). `$_SESSION[\u0027user\u0027]` is set, `setUserCookie` runs, `_session_regenerate_id` issues a new session ID, and the response carries `Set-Cookie: PHPSESSID=\u003cnew-sid\u003e; ...`.\n   - Subsequent code runs the encoder pipeline as admin \u2014 but the attacker\u0027s primary goal was already achieved when the session was established.\n4. Attacker captures the `Set-Cookie: PHPSESSID=...` header from the response and uses that cookie on all subsequent requests. Server treats them as user 1 (admin) \u2014 full UI access, all admin endpoints, all video management, plugin configuration, user impersonation, etc.\n5. Final state: admin account takeover. The original Meet recorder\u0027s flow (legitimate uploads with `users_id` = the user who scheduled the meeting) is indistinguishable on the wire from the attack flow (`users_id` = whoever the attacker wants to be).\n\n## Security Impact\n\n**Severity:** sec-high. End state is full account takeover of any user (including admin), reachable from a single HTTP POST once the secret is known. The shared-secret precondition raises AC to High but does not eliminate it as a credible threat \u2014 the secret is computable from any leak of `videos/configuration.php`, and AVideo\u0027s CVE history in that surface area is non-trivial.\n**Attacker capability:** session hijack as any `users_id` the attacker cares to name. The attacker chooses the target by setting the filename\u0027s leading digits before the first `-`. No bound on which user IDs are reachable; admin (`1` on a default install) is the obvious target. Once the session is captured, the attacker has full admin UI/API access for the session lifetime (hours-to-days depending on `rememberme` flag).\n**Preconditions:** Meet plugin enabled (default-off but commonly enabled by deployments using AVideo for video-conferencing recording). Knowledge of the Meet shared secret (computable from the salt; obtainable via timing attack on `checkToken.json.php`).\n**Differential:** source-inspection-verified end-to-end. The two relevant code blocks are quoted verbatim in \u00a7Affected Code; both lines are reachable on every successful POST to the endpoint. The patched build (with the suggested fix below) either rejects the upload as `\u0027cannot derive identity from filename\u0027` or constrains the `users_id` to one bound by an additional signed claim from the Meet recorder.\n\n## Suggested Fix\n\nThree changes, in order of importance:\n\n```diff\n--- a/plugin/Meet/uploadRecordedVideo.json.php\n+++ b/plugin/Meet/uploadRecordedVideo.json.php\n@@ -53,17 +53,28 @@ if (empty($_FILES[\u0027upl\u0027])) {\n     forbiddenPage(\u0027videoFile not found\u0027);\n }\n\n-$users_id = explode(\u0027-\u0027, $_FILES[\u0027upl\u0027][\u0027name\u0027])[0];\n+// The users_id MUST come from a signed claim (e.g., a JWT issued by AVideo\n+// when the meeting was scheduled), not from a filename the caller controls.\n+// Verify a recording-upload token here that was minted at meeting-create\n+// time and bound to (meet_schedule_id, users_id) with an HMAC.\n+$claim = MeetUploadClaim::verifyFromHeaders($headers);\n+if (!$claim) {\n+    forbiddenPage(\u0027Missing or invalid recording upload claim\u0027);\n+}\n+$users_id = (int) $claim-\u003eusers_id;\n+if (!$users_id || !User::idExists($users_id)) {\n+    forbiddenPage(\u0027Recording upload claim references unknown user\u0027);\n+}\n\n-$userObject = new User($users_id);\n-$userObject-\u003elogin(true, true);\n+// Credit the upload to $users_id WITHOUT establishing a session. The encoder\n+// pipeline can be parameterised to record ownership directly; there is no\n+// reason for a service-to-service upload endpoint to mint a user session.\n+$queueOwnerUsersId = $users_id;\n $tmpFile = getTmpDir() . uniqid();\n\n if (move_uploaded_file($_FILES[\u0027upl\u0027][\u0027tmp_name\u0027], $tmpFile)) {\n     $_FILES[\u0027upl\u0027][\u0027tmp_name\u0027] = $tmpFile;\n-    require $global[\u0027systemRootPath\u0027] . \u0027objects/aVideoQueueEncoder.json.php\u0027;\n+    aVideoQueueEncoder::encodeOnBehalfOf($queueOwnerUsersId, $_FILES[\u0027upl\u0027]);\n }\n```\n\nAdditionally:\n\n1. **Use `hash_equals` for the secret comparison** in both this endpoint and `checkToken.json.php` (`if (!hash_equals($objM-\u003esecret, $token))`). The current `==`/`===` is vulnerable to byte-by-byte timing analysis.\n2. **Remove `checkToken.json.php` entirely**, or at least gate it behind `User::isAdmin()`. A network-reachable endpoint that confirms whether a guess matches the server-side secret is exactly the wrong shape for a high-value secret like this one.\n\nOptional defense-in-depth (separate change): rotate the Meet secret to use a random 256-bit value (not derived from `salt`), so a `videos/configuration.php` disclosure does not also yield the Meet secret. Store the random secret as a per-deployment row in the Meet plugin\u0027s configuration table, generated at first-run.\n\nAdd a regression test: call `uploadRecordedVideo.json.php` with the correct secret but a filename of `1-x.mp4`; assert the response does NOT include a `Set-Cookie: PHPSESSID=` header.",
  "id": "GHSA-qxvm-r42f-5p8j",
  "modified": "2026-05-15T18:17:19Z",
  "published": "2026-05-15T18:17:19Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/WWBN/AVideo/security/advisories/GHSA-qxvm-r42f-5p8j"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/WWBN/AVideo"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "AVideo\u0027s Meet plugin: `uploadRecordedVideo.json.php` derives `users_id` from the uploaded filename and calls passwordless `User-\u003elogin()`, allowing any caller with the Meet shared secret to obtain a session as arbitrary users including admin"
}
Mitigation
Architecture and Design

For each and every data access, ensure that the user has sufficient privilege to access the record that is being requested.

Mitigation
Architecture and Design Implementation

Make sure that the key that is used in the lookup of a specific user's record is not controllable externally by the user or that any tampering can be detected.

Mitigation
Architecture and Design

Use encryption in order to make it more difficult to guess other legitimate values of the key or associate a digital signature with the key so that the server can verify that there has been no tampering.

No CAPEC attack patterns related to this CWE.

← Back to the CWE list