GHSA-524M-Q5M7-79MM
Vulnerability from github – Published: 2026-01-13 15:11 – Updated: 2026-01-13 15:11Summary The Mailpit WebSocket server is configured to accept connections from any origin. This lack of Origin header validation introduces a Cross-Site WebSocket Hijacking (CSWSH) vulnerability.
An attacker can host a malicious website that, when visited by a developer running Mailpit locally, establishes a WebSocket connection to the victim's Mailpit instance (default ws://localhost:8025). This allows the attacker to intercept sensitive data such as email contents, headers, and server statistics in real-time.
Vulnerable Code The vulnerability exists in server/websockets/client.go where the CheckOrigin function is explicitly set to return true for all requests, bypassing standard Same-Origin Policy (SOP) protections provided by the gorilla/websocket library.
https://github.com/axllent/mailpit/blob/877a9159ceeaf380d5bb0e1d84017b24d2e7b361/server/websockets/client.go#L34-L39
Impact This vulnerability impacts the Confidentiality of the data stored in or processed by Mailpit. Although Mailpit is often used as a local development tool, this vulnerability allows remote exploitation via a web browser.
- Scenario: A developer has Mailpit running at localhost:8025.
- Trigger: The developer visits a malicious website (or a compromised legitimate site) in the same browser.
- Exploitation: The malicious site's JavaScript initiates a WebSocket connection to ws://localhost:8025/api/events. Since the origin check is disabled, the browser allows this cross-origin connection.
- Data Leak: The attacker receives all broadcasted events, including full email details (subjects, sender/receiver info) and server metrics.
Attack Impact - Real-time notification of new emails - Email metadata (sender, subject, recipients) - Mailbox statistics - All WebSocket broadcast data
Recommended Fix
The CheckOrigin function should be removed to allow gorilla/websocket to enforce its default safe behavior (checking that the Origin matches the Host). Alternatively, strict validation logic should be implemented.
Proposed Change (Remove unsafe check):
var upgrader = websocket.Upgrader{
ReadBufferSize: 1024,
WriteBufferSize: 1024,
// CheckOrigin: func(r *http.Request) bool { return true }, // REMOVED
EnableCompression: true,
}
Proof of Concept (PoC): To reproduce this vulnerability:
- Start Mailpit (default settings).
- Save the following HTML code as poc.html and serve it from a different origin (e.g., using python http.server on port 8000 or opening it directly as a file).
- Open the poc_websocket_hijack.html file in your browser.
- Send a test email to Mailpit or perform any action in the Mailpit UI.
- Observe that the "malicious" page successfully receives the event data.
{
"affected": [
{
"package": {
"ecosystem": "Go",
"name": "github.com/axllent/mailpit"
},
"ranges": [
{
"events": [
{
"introduced": "1.2.6"
},
{
"fixed": "1.28.2"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Go",
"name": "github.com/axllent/mailpit"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "0.0.0-20260110031614"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-22689"
],
"database_specific": {
"cwe_ids": [
"CWE-1385"
],
"github_reviewed": true,
"github_reviewed_at": "2026-01-13T15:11:42Z",
"nvd_published_at": "2026-01-10T06:15:51Z",
"severity": "MODERATE"
},
"details": "**Summary**\nThe Mailpit WebSocket server is configured to accept connections from any origin. This lack of Origin header validation introduces a Cross-Site WebSocket Hijacking (CSWSH) vulnerability.\n\nAn attacker can host a malicious website that, when visited by a developer running Mailpit locally, establishes a WebSocket connection to the victim\u0027s Mailpit instance (default ws://localhost:8025). This allows the attacker to intercept sensitive data such as email contents, headers, and server statistics in real-time.\n\n**Vulnerable Code**\nThe vulnerability exists in server/websockets/client.go where the CheckOrigin function is explicitly set to return true for all requests, bypassing standard Same-Origin Policy (SOP) protections provided by the gorilla/websocket library.\n\nhttps://github.com/axllent/mailpit/blob/877a9159ceeaf380d5bb0e1d84017b24d2e7b361/server/websockets/client.go#L34-L39\n\n**Impact**\nThis vulnerability impacts the Confidentiality of the data stored in or processed by Mailpit.\nAlthough Mailpit is often used as a local development tool, this vulnerability allows remote exploitation via a web browser.\n\n- **Scenario**: A developer has Mailpit running at localhost:8025.\n- **Trigger**: The developer visits a malicious website (or a compromised legitimate site) in the same browser.\n- **Exploitation**: The malicious site\u0027s JavaScript initiates a WebSocket connection to ws://localhost:8025/api/events. Since the origin check is disabled, the browser allows this cross-origin connection.\n- **Data Leak**: The attacker receives all broadcasted events, including full email details (subjects, sender/receiver info) and server metrics.\n\n**Attack Impact**\n- Real-time notification of new emails\n- Email metadata (sender, subject, recipients)\n- Mailbox statistics\n- All WebSocket broadcast data\n\n**Recommended Fix**\nThe `CheckOrigin` function should be removed to allow gorilla/websocket to enforce its default safe behavior (checking that the Origin matches the Host). Alternatively, strict validation logic should be implemented.\n\n**Proposed Change (Remove unsafe check):**\n\n```go\nvar upgrader = websocket.Upgrader{\n ReadBufferSize: 1024,\n WriteBufferSize: 1024,\n // CheckOrigin: func(r *http.Request) bool { return true }, // REMOVED\n EnableCompression: true,\n}\n```\n\n**Proof of Concept (PoC)**: To reproduce this vulnerability:\n\n- Start Mailpit (default settings).\n- Save the following HTML code as poc.html and serve it from a different origin (e.g., using python http.server on port 8000 or opening it directly as a file).\n- Open the [poc_websocket_hijack.html](https://github.com/user-attachments/files/24522726/poc_websocket_hijack.html) file in your browser.\n- Send a test email to Mailpit or perform any action in the Mailpit UI.\n- Observe that the \"malicious\" page successfully receives the event data.",
"id": "GHSA-524m-q5m7-79mm",
"modified": "2026-01-13T15:11:42Z",
"published": "2026-01-13T15:11:42Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/axllent/mailpit/security/advisories/GHSA-524m-q5m7-79mm"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-22689"
},
{
"type": "WEB",
"url": "https://github.com/axllent/mailpit/commit/6f1f4f34c98989fd873261018fb73830b30aec3f"
},
{
"type": "PACKAGE",
"url": "https://github.com/axllent/mailpit"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N",
"type": "CVSS_V3"
}
],
"summary": "Mailpit is vulnerable to Cross-Site WebSocket Hijacking (CSWSH) allowing unauthenticated access to emails"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.