Description

FreeScout is a self-hosted help desk and shared mailbox application. Prior to version 1.8.226, the  polycast/receive  endpoint fails to escape SQL LIKE wildcards in supplied channel names, while channel authorization applies only to names with specific prefixes. An unauthenticated attacker with a guest session and valid CSRF token can submit a wildcard channel name to retrieve matching notification events intended for other users. Notification payload generation lacks a corresponding access check, exposing conversation message content, identifiers, URLs, and notification HTML across users and mailboxes. This issue is fixed in version 1.8.226.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "7.7",
  "pubDate": "2026-11-06T17:18:34.047Z",
  "pubdate": "2026-11-06T17:18:34.047Z",
  "executiveSummary": "FreeScout versions prior to 1.8.226 contain a critical vulnerability in the polycast/receive endpoint, resulting from improper sanitization of SQL LIKE wildcards within channel name parameters.\nThis vulnerability allows an unauthenticated attacker, utilizing a guest session and a valid CSRF token, to bypass authorization controls through wildcard manipulation.\nBy submitting specifically crafted channel names, an attacker can retrieve notification events and sensitive conversation data—including message content, identifiers, and HTML payloads—that belong to other users and mailboxes.\nThe risk implication is significant, as the application fails to perform necessary access checks during the notification payload generation process, leading to a complete compromise of communication privacy across multiple mailboxes.\nSuccessful exploitation requires the attacker to possess a valid CSRF token, which is obtainable via a guest session, but does not require higher-level authentication. Organizations running FreeScout versions older than 1.8.226 are at risk of unauthorized data exfiltration and should prioritize updating their software immediately.",
  "technicalDetails": "The vulnerability resides within the polycast/receive endpoint of the FreeScout application, which is responsible for handling incoming notification events. The root cause is the failure to escape SQL LIKE wildcard characters (%) within the input provided for channel names.\nThe application implements an authorization mechanism that filters access based on specific channel name prefixes; however, because the backend database query uses the unescaped channel name in a LIKE clause, an attacker can supply wildcard characters to match unintended channel strings that would otherwise be protected or restricted.\nThe exploitation flow initiates when an attacker obtains a valid CSRF token, which is accessible via an unauthenticated guest session. The attacker then submits a malicious request to the polycast/receive endpoint, providing a channel name argument containing wildcard patterns (e.g., '%').\nBecause the database query is not properly sanitized, the SQL engine interprets these wildcards, causing the application to return notification records matching the attacker-supplied pattern, effectively bypassing the intended prefix-based authorization constraints.\nFollowing the unauthorized retrieval of these notification objects, the application proceeds to generate the notification payload. Crucially, the system fails to perform any secondary access control checks during this generation phase. Consequently, the application processes and returns the full notification data, which includes highly sensitive information such as internal conversation message bodies, unique identifiers, conversation URLs, and pre-rendered notification HTML content.\nThis behavior exposes inter-mailbox data to any user capable of initiating the request, allowing an attacker to systematically scrape communication metadata and message content across the entire application instance. The impact is a total breakdown of multi-tenancy and mailbox isolation. The vulnerability persists in all FreeScout versions prior to 1.8.226, where the logic was corrected to strictly enforce input sanitization and access control validation."
}