CVE-2026-86345 in Directory Server
Summary
by MITRE • 10/02/2026
A flaw was found in 389-ds-base. The server does not discard plaintext bytes already buffered from a client connection when negotiating StartTLS, allowing an on-path attacker to inject a crafted LDAP message that is processed after the TLS upgrade and whose response is delivered to the client in place of the client's own pending operation's response, due to messageID collision. This can cause a client application to treat a failed authentication (bind) attempt as successful.
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 10/02/2026
The vulnerability identified within 389-ds-base represents a critical flaw in the implementation of the StartTLS protocol negotiation process for LDAP connections. During the establishment of an encrypted channel, the server fails to properly discard plaintext bytes that have already been buffered from the client connection before the TLS handshake is fully completed and encryption begins. This architectural oversight creates a window where data sent prior to the upgrade remains accessible in the buffer rather than being cleared or ignored as required by secure protocol standards. The core technical issue lies in how the server handles message sequencing during this transition period, specifically regarding the management of pending operations and their associated identifiers.
An on-path attacker positioned between the client and the 389 Directory Server can exploit this buffering behavior to inject a crafted LDAP message into the stream. Because the plaintext bytes are not discarded, the attacker can manipulate the timing or content such that an injected message shares the same messageID as a pending operation initiated by the legitimate client. In LDAP protocols, the messageID is used to correlate requests with their corresponding responses. When this collision occurs, the server processes the injected request and generates a response based on it. Due to the matching identifier, this forged response is delivered back to the client application instead of the actual result of the client's original operation.
The operational impact of this vulnerability is severe, particularly concerning authentication mechanisms. A primary consequence described in the analysis is that a client application may interpret a failed authentication attempt as successful. For instance, if an attacker injects a bind request with valid credentials or manipulates the response to indicate success while the actual user's credentials were invalid, the server returns a positive acknowledgment for the forged operation. The client, relying on the messageID correlation, accepts this response as its own outcome. This effectively bypasses authentication controls, allowing unauthorized access to directory services and sensitive data stored within them.
This flaw aligns with CWE-362, which describes concurrent execution using shared resources with improper synchronization, specifically in the context of race conditions or state management errors during protocol transitions. Furthermore, it relates to CWE-489, Active Debug Code, if one considers the retention of sensitive plaintext data as a form of unintended information exposure that facilitates further attacks. From an offensive security perspective, this technique is consistent with MITRE ATT&CK tactic T1076, Remote Protocol Impersonation, where an attacker intercepts and manipulates network traffic to impersonate legitimate users or services. The specific method of exploiting messageID collisions falls under techniques designed to bypass authentication mechanisms by manipulating protocol state.
Mitigation strategies must focus on ensuring strict separation between plaintext and encrypted data streams during the StartTLS upgrade process. Administrators should verify that their version of 389-ds-base includes patches addressing this buffering issue, as upstream developers have likely implemented fixes to clear buffers or enforce stricter messageID isolation before encryption is established. Additionally, implementing certificate pinning on client applications can reduce the risk of successful man-in-the-middle attacks by ensuring clients only trust specific server identities. Network-level monitoring and intrusion detection systems should also be configured to flag anomalous LDAP traffic patterns that may indicate such injection attempts, although patching remains the primary defense against this protocol-level flaw.