HIGH 7.5 PyPI

mail-parser: Sender IP address spoofing in get_server_ipaddress()

GHSA-72px-j8gh-p79x · CVE-2026-129297

Published · Modified

AI SAST

Find this class of vulnerability in your own code

Corgea's AI-native static analysis detects vulnerabilities like this one across your repositories, ranks them by exploitability, and returns review-ready fixes.

Description

Patches

Fixed in 4.6.2. Upgrade with pip install --upgrade mail-parser.

Candidate addresses are now selected by position, not by appearance — a bracketed address
the sender wrote is byte-identical to one the MTA wrote, so no amount of pattern matching
can tell them apart:

  • the first token of the from clause is the HELO name, whatever its shape, and is never a
    candidate;
  • an explicit HELO marker inside a comment group is sender text and is excluded;
  • a candidate must sit inside a ( / [ group;
  • IPv4 and IPv6 matches are merged positionally, never family-first;
  • the walk stops at the first trusted hop naming no address, rather than continuing into
    sender-written headers.

The IPv6 pattern was corrected to skip the IPv6: tag and to prefer the longest valid form.

Behaviour change. get_server_ipaddress() now returns None rather than an unreliable
value in two cases that previously returned an address: a from clause consisting of a bare
address with no HELO marker (indistinguishable from EHLO <that address>), and a trusted hop
naming no address at all. Callers must handle None.

Workarounds

None within mail-parser. Callers who cannot upgrade should not treat the return value of
get_server_ipaddress() as trustworthy, and should derive sender attribution from an
authenticated signal instead — the MTA's own logs, or SPF/DKIM/DMARC results computed
outside this library.

Ready to move

Start Securing

Free, no credit card | First findings in minutes