GHSA-8CGQ-X343-P74C
Vulnerability from github – Published: 2026-11-16 22:22 – Updated: 2026-11-16 22:22Impact
write_attachments() called _safe_attachment_filename() and write_sample() inside its loop
with no guard, so a single crafted attachment raised out of the whole save. Every attachment
after the malicious one was silently never written, with no error surfaced to a caller that
catches only MailParserError.
That is an evasion primitive as much as an availability bug: a sender puts one poisoned part first and the real payload second, and the payload never reaches the extraction directory that a downstream scanner or sandbox reads — while it still reaches the recipient's mail client intact. The tool's output diverges from the bytes actually in the message, which is the same failure mode as the filename-sanitizing bypass in GHSA-fr3h-cpq2-9496.
Three shapes were reachable, all from attacker-controlled MIME metadata:
- NUL smuggled through RFC 2231 —
filename*=us-ascii''evil%00.bindecodes to a NUL in the filename during parsing and raisedValueErrorout of the name check; - overlong basename — a filename longer than
NAME_MAX(e.g. 300 characters) raisedOSError: File name too longfromopen(), after creating a zero-byte stub on disk; - undecodable payload — a body that is not valid base64 raised
binascii.Error(aValueErrorsubclass) from the decode performed inside theopen()block.
import tempfile, mailparser
raw = ("Content-Type: multipart/mixed; boundary=b\r\n\r\n"
"--b\r\nContent-Type: application/octet-stream\r\n"
"Content-Disposition: attachment; filename*=us-ascii''evil%00.bin\r\n\r\nxx\r\n"
"--b\r\nContent-Type: application/octet-stream\r\n"
"Content-Disposition: attachment; filename=good.bin\r\n\r\nyy\r\n"
"--b--\r\n")
mp = mailparser.parse_from_string(raw) # parse itself is fine; safe_filename is None
mp.write_attachments(tempfile.mkdtemp())
# 4.6.0 - 4.6.2 -> ValueError, and good.bin is never written
# 4.6.3 -> logs "Skipped attachment with invalid filename: 'evil\x00.bin'";
# good.bin is written
parse() itself was never affected — it catches the ValueError — so only the write path was
unguarded. The reject-paths that made this reachable were introduced with the filename
sanitizing in 4.6.0 (PR #165).
Patches
Fixed in 4.6.3. write_attachments() now logs and skips a single unusable attachment
instead of aborting the batch:
- an invalid filename is caught and skipped;
- overlong basenames are truncated to 240 UTF-8 bytes, preserving the extension, rather than
reaching
open(); - the base64 decode moved before
open(), so a failure leaves no zero-byte stub, and it now repairs padding and alphabet errors the way mail clients do (matchingMIME::Base64on 2008 cases) — rejecting them let a sender append one character to make an attachment vanish from the extraction directory while it still reached the recipient; MailParserPathErroris deliberately not caught by the per-attachment handler: a containment failure is not a per-attachment problem, and path-containment checks now raiseMailParserPathErrorinstead ofValueErrorso the broader catch cannot swallow them.
Deduplication was hardened in the same change: it reserves room for its own suffix marker
(write_sample() sanitizes again, and the clamp there used to cut the marker back off,
collapsing distinct attachments onto one file), resumes from the last used suffix per stem
instead of rescanning from 1 (quadratic before — 16000 same-named attachments took 14 s, now
milliseconds), and folds case, because APFS, exFAT and SMB collapse Invoice.pdf and
invoice.pdf.
Workarounds
Upgrade to 4.6.3. If you cannot upgrade, do not call write_attachments() on untrusted mail;
iterate parser.attachments yourself and write each attachment inside its own
try/except Exception, skipping the ones whose safe_filename is None.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "mail-parser"
},
"ranges": [
{
"events": [
{
"introduced": "4.6.0"
},
{
"fixed": "4.6.3"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-129301"
],
"database_specific": {
"cwe_ids": [
"CWE-248",
"CWE-436"
],
"github_reviewed": true,
"github_reviewed_at": "2026-11-16T22:22:13Z",
"nvd_published_at": "2026-11-16T18:17:31Z",
"severity": "MODERATE"
},
"details": "### Impact\n\n`write_attachments()` called `_safe_attachment_filename()` and `write_sample()` inside its loop\nwith no guard, so a single crafted attachment raised out of the whole save. Every attachment\n**after** the malicious one was silently never written, with no error surfaced to a caller that\ncatches only `MailParserError`.\n\nThat is an evasion primitive as much as an availability bug: a sender puts one poisoned part\nfirst and the real payload second, and the payload never reaches the extraction directory that\na downstream scanner or sandbox reads \u2014 while it still reaches the recipient\u0027s mail client\nintact. The tool\u0027s output diverges from the bytes actually in the message, which is the same\nfailure mode as the filename-sanitizing bypass in\n[GHSA-fr3h-cpq2-9496](https://github.com/SpamScope/mail-parser/security/advisories/GHSA-fr3h-cpq2-9496).\n\nThree shapes were reachable, all from attacker-controlled MIME metadata:\n\n- **NUL smuggled through RFC 2231** \u2014 `filename*=us-ascii\u0027\u0027evil%00.bin` decodes to a NUL in the\n filename during parsing and raised `ValueError` out of the name check;\n- **overlong basename** \u2014 a filename longer than `NAME_MAX` (e.g. 300 characters) raised\n `OSError: File name too long` from `open()`, *after* creating a zero-byte stub on disk;\n- **undecodable payload** \u2014 a body that is not valid base64 raised `binascii.Error` (a\n `ValueError` subclass) from the decode performed inside the `open()` block.\n\n```python\nimport tempfile, mailparser\n\nraw = (\"Content-Type: multipart/mixed; boundary=b\\r\\n\\r\\n\"\n \"--b\\r\\nContent-Type: application/octet-stream\\r\\n\"\n \"Content-Disposition: attachment; filename*=us-ascii\u0027\u0027evil%00.bin\\r\\n\\r\\nxx\\r\\n\"\n \"--b\\r\\nContent-Type: application/octet-stream\\r\\n\"\n \"Content-Disposition: attachment; filename=good.bin\\r\\n\\r\\nyy\\r\\n\"\n \"--b--\\r\\n\")\n\nmp = mailparser.parse_from_string(raw) # parse itself is fine; safe_filename is None\nmp.write_attachments(tempfile.mkdtemp())\n# 4.6.0 - 4.6.2 -\u003e ValueError, and good.bin is never written\n# 4.6.3 -\u003e logs \"Skipped attachment with invalid filename: \u0027evil\\x00.bin\u0027\";\n# good.bin is written\n```\n\n`parse()` itself was never affected \u2014 it catches the `ValueError` \u2014 so only the write path was\nunguarded. The reject-paths that made this reachable were introduced with the filename\nsanitizing in 4.6.0 (PR #165).\n\n### Patches\n\nFixed in **4.6.3**. `write_attachments()` now logs and skips a single unusable attachment\ninstead of aborting the batch:\n\n- an invalid filename is caught and skipped;\n- overlong basenames are truncated to 240 UTF-8 bytes, preserving the extension, rather than\n reaching `open()`;\n- the base64 decode moved *before* `open()`, so a failure leaves no zero-byte stub, and it now\n repairs padding and alphabet errors the way mail clients do (matching `MIME::Base64` on 2008\n cases) \u2014 rejecting them let a sender append one character to make an attachment vanish from\n the extraction directory while it still reached the recipient;\n- `MailParserPathError` is deliberately **not** caught by the per-attachment handler: a\n containment failure is not a per-attachment problem, and path-containment checks now raise\n `MailParserPathError` instead of `ValueError` so the broader catch cannot swallow them.\n\nDeduplication was hardened in the same change: it reserves room for its own suffix marker\n(`write_sample()` sanitizes again, and the clamp there used to cut the marker back off,\ncollapsing distinct attachments onto one file), resumes from the last used suffix per stem\ninstead of rescanning from 1 (quadratic before \u2014 16000 same-named attachments took 14 s, now\nmilliseconds), and folds case, because APFS, exFAT and SMB collapse `Invoice.pdf` and\n`invoice.pdf`.\n\n### Workarounds\n\nUpgrade to 4.6.3. If you cannot upgrade, do not call `write_attachments()` on untrusted mail;\niterate `parser.attachments` yourself and write each attachment inside its own\n`try`/`except Exception`, skipping the ones whose `safe_filename` is `None`.",
"id": "GHSA-8cgq-x343-p74c",
"modified": "2026-11-16T22:22:13Z",
"published": "2026-11-16T22:22:13Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/SpamScope/mail-parser/security/advisories/GHSA-8cgq-x343-p74c"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-129301"
},
{
"type": "WEB",
"url": "https://github.com/SpamScope/mail-parser/pull/171"
},
{
"type": "WEB",
"url": "https://github.com/SpamScope/mail-parser/commit/11b894a0b8b3f11055a99caff118206138d1daba"
},
{
"type": "WEB",
"url": "https://github.com/SpamScope/mail-parser/commit/e1d9fe6778045d6e2c44406bc8822e3464c1097f"
},
{
"type": "PACKAGE",
"url": "https://github.com/SpamScope/mail-parser"
},
{
"type": "WEB",
"url": "https://github.com/SpamScope/mail-parser/releases/tag/4.6.3"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:L",
"type": "CVSS_V3"
}
],
"summary": "mail-parser: One hostile attachment aborts `write_attachments()` and silently drops the rest of the batch"
}
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.