Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-107576 Overview
CVE-2026-107576 is an algorithmic complexity vulnerability [CWE-407] in Progressive Robot hMailServer versions 6.0.0 through 6.3.5. The flaw resides in the inbound DomainKeys Identified Mail (DKIM) and Authenticated Received Chain (ARC) signature verification routines. A remote, unauthenticated attacker can send a single crafted message that keeps worker threads busy for tens of seconds per signature. Because hMailServer evaluates up to ten signatures per message during both DKIM and Domain-based Message Authentication, Reporting and Conformance (DMARC) checks on delivery and Simple Mail Transfer Protocol (SMTP) threads, the attacker can exhaust mail service availability.
Critical Impact
A single inbound message with a crafted header can stall hMailServer delivery and SMTP worker threads, rendering mail services unavailable to all users.
Affected Products
- Progressive Robot hMailServer 6.0.0 through 6.3.5
- Any deployment accepting inbound SMTP with DKIM verification enabled
- Any deployment performing DMARC evaluation on inbound mail
Discovery Timeline
- 2026-10-08 - CVE-2026-107576 published to the National Vulnerability Database
- 2026-10-08 - Last updated in NVD database
Technical Details for CVE-2026-107576
Vulnerability Analysis
The vulnerability stems from quadratic time complexity in two cooperating code paths used by DKIM and ARC signature verification. The first path builds the canonical header. Under simple canonicalisation, each continuation line of a folded field is prepended to lines already gathered, repeating prior work on every iteration. The second path selects header fields named in a signature's h= tag. For each name, the implementation searches the collected field list from the bottom and erases the matched entry from the middle of the list, forcing repeated list traversal and shifting.
Both operations grow with the square of the header size. A message that contains many header fields, or a single field folded across many lines, forces a verification worker to spend tens of seconds per signature. hMailServer evaluates up to ten DKIM signatures per message, and DMARC evaluation repeats signature processing. These costs accumulate on the threads that handle SMTP ingress and message delivery.
Root Cause
The root cause is inefficient algorithmic complexity in header canonicalisation and header-field selection. The code chooses data structures and iteration patterns that convert linear input growth into quadratic processing time, without any upper bound on header count or folding depth.
Attack Vector
Exploitation requires no authentication and no user interaction. The attacker publishes a DKIM key for a domain they control, signs a message with that domain's DKIM-Signature, and sends the message to any address accepted by the target hMailServer. The crafted message carries an oversized header, either through a very large number of fields or through a single deeply folded field. Repeating this across concurrent connections saturates SMTP and delivery worker threads, denying mail service to legitimate users. Verified code examples are not available; refer to the GitLab Work Item #73 for upstream technical discussion.
Detection Methods for CVE-2026-107576
Indicators of Compromise
- Inbound SMTP messages containing hundreds or thousands of header fields, or header fields folded across many continuation lines
- DKIM-Signature headers referencing attacker-controlled domains with abnormally long h= tag lists
- Prolonged CPU utilization on hMailServer worker threads during inbound message processing
- SMTP connection timeouts and queue backlog growth coinciding with specific inbound messages
Detection Strategies
- Inspect SMTP ingress logs for messages where DKIM or DMARC verification duration exceeds normal baselines by an order of magnitude
- Alert on inbound messages whose header section exceeds reasonable size thresholds, such as more than a few hundred header fields or more than a few thousand folded continuation lines
- Correlate sudden hMailServer process CPU spikes with the message IDs being processed at that moment
Monitoring Recommendations
- Monitor hMailServer worker thread count, queue depth, and SMTP session duration for anomalies
- Track per-sender message rates and reject or rate-limit senders whose messages consistently trigger long verification times
- Capture and retain full headers of rejected or delayed messages for forensic review
How to Mitigate CVE-2026-107576
Immediate Actions Required
- Upgrade hMailServer to version 6.3.6 or later, which addresses the quadratic complexity in DKIM and ARC verification
- Enforce inbound message size limits and header count limits at the perimeter mail gateway
- Rate-limit inbound SMTP connections per source IP to blunt concurrent exploitation attempts
- Review mail server logs for prior instances of long-duration DKIM verification
Patch Information
Progressive Robot released hMailServer 6.3.6 to remediate CVE-2026-107576. The release notes and source changes are available on the project's GitLab page: GitLab Release v6.3.6. The upstream tracking item is published as GitLab Work Item #73.
Workarounds
- Place a mail security gateway in front of hMailServer that performs DKIM and DMARC verification with bounded header processing
- Temporarily disable inbound DKIM and ARC verification on hMailServer if an upstream gateway already enforces equivalent checks
- Reject inbound messages with abnormally large headers at the SMTP layer before they reach verification routines
# Example perimeter guardrail: cap inbound header size at a mail gateway
# Postfix front-end example (header_size_limit in bytes)
header_size_limit = 102400
message_size_limit = 10485760
smtpd_client_connection_rate_limit = 30
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

