#1136004 hashcat: CVE-2026-42482 CVE-2026-42483 CVE-2026-42484

Package:
src:hashcat
Source:
src:hashcat
Submitter:
Moritz Mühlenhoff
Date:
2026-11-16 20:25:02 UTC
Severity:
normal
Tags:
#1136004#5
Date:
2026-05-08 13:20:22 UTC
From:
To:
Hi,

The following vulnerabilities were published for hashcat.

CVE-2026-42482[0]:
| A stack-based buffer overflow in mangle_to_hex_lower() and
| mangle_to_hex_upper() in src/rp_cpu.c in hashcat v7.1.2 allows an
| attacker to cause a denial of service or possibly execute arbitrary
| code via a crafted rule file, or via the -j or -k rule options used
| with password candidates of 128 or more characters. The
| vulnerability is caused by a bounds check that fails to account for
| the 2x expansion that occurs when password bytes are converted to
| hexadecimal.

CVE-2026-42483[1]:
| A heap-based buffer overflow in the Kerberos hash parser in hashcat
| v7.1.2 allows an attacker to cause a denial of service or possibly
| execute arbitrary code via a crafted Kerberos hash file. The issue
| affects module_hash_decode in multiple Kerberos-related modules
| because account_info_len is calculated from untrusted delimiter
| positions without upper-bound validation before memcpy copies the
| data into a fixed-size account_info buffer.

CVE-2026-42484[2]:
| A heap-based buffer overflow in hex_to_binary in the PKZIP hash
| parser in hashcat v7.1.2 allows an attacker to cause a denial of
| service or possibly execute arbitrary code via a crafted PKZIP hash
| file. The issue affects modules 17200, 17210, 17220, 17225, and
| 17230. When data_type_enum<=1, attacker-controlled hex data from a
| user-supplied hash string is decoded into a fixed-size buffer
| without proper input-length validation.

It's unclear whether this has been properly reported upstream:
https://gist.github.com/sgInnora/107f2eb20367e47d58c911e38d56a91f



If you fix the vulnerabilities please also make sure to include the
CVE (Common Vulnerabilities & Exposures) ids in your changelog entry.

For further information see:

[0] https://security-tracker.debian.org/tracker/CVE-2026-42482
https://www.cve.org/CVERecord?id=CVE-2026-42482
[1] https://security-tracker.debian.org/tracker/CVE-2026-42483
https://www.cve.org/CVERecord?id=CVE-2026-42483
[2] https://security-tracker.debian.org/tracker/CVE-2026-42484
https://www.cve.org/CVERecord?id=CVE-2026-42484

Please adjust the affected versions in the BTS as needed.

#1136004#12
Date:
2026-11-03 11:58:56 UTC
From:
To:
Hi,

I mapped the three CVEs against upstream hashcat and Debian's
7.1.2+ds1-4. All are fixed in upstream v8.0.0 (released 2026-10-25).

CVE-2026-42482 (mangle_to_hex_lower/upper in src/rp_cpu.c)
Debian 7.1.2 does NOT seem to be affected. These functions do not
exist in 7.1.2: neither the v7.1.2 tag nor the 7.1.2+ds1-4 source
(
https://www.google.com/url?q=http://sources.debian.org&source=gmail&ust=1793793536257000&sa=E)
contains "to_hex" in src/rp_cpu.c. The hex rule
ops were added later (195e14cb82 "Initial hex rule support",
2025-10-09). The reporter's write-up says "v7.1.2" but quotes code
that is only on later master, which also explains why it lists the
Kerberos modules 35300/35400; those don't exist in 7.1.2 either.
The fix upstream is 5381eddb13 "Fix buffer overflow in rule parsing
(#4618)", which is the PR the write-up refers to. Please double-check,
but I believe this CVE can be marked not-affected for 7.1.2.

CVE-2026-42484 (PKZIP hex_to_binary heap overflow; modes 17200, 17210,
17220, 17225, 17230)
Fix: 9468ae18f2761a96c3f1c3247f3ca184c4e9ac22
"Fix PKZIP module heap overflow causing crashes (#4690)"
It bounds data_length and the hex length (<= MAX_DATA*2, even). It
applies cleanly to 7.1.2.

CVE-2026-42483 (Kerberos account_info / domain overflows)
The write-up covers two different bugs:
a) account_info length unbounded (13100, 18200):
caa6563ef9dbd347f1c935212912cafd1f8a37cf "Fix 0x80 padding byte
overflow causing crashes/corrupt output (#4698)", tightened from >
to >= in 836f11de165a1703d772fbafd82e17422b56f029.
b) etype 17/18 modules (19600, 19700, 19800, 19900, 28800, 28900,
32100, 32200): the struct fields are bounded by the tokenizer
(len_max 512), but the parsers copy the domain into a local
"u8 domain[128]" stack buffer:
6a3994efda07d7200511de2f5382e724de6a85a4
"Fix Kerberos etype 17/18 domain buffer overflow (#4689)"
It changes the local buffer to domain[512].

Backport for 7.1.2 (attached, against the upstream v7.1.2 tag;
src/modules only):
- 9468ae18f2 applies as is.
- The decode-side hunks of caa6563ef9 and 6a3994efda apply as is. Their
module_hash_encode hunks don't, because 7.1.2 uses stack arrays where
upstream later moved to hcmalloc(). I applied the same "+1 for the
NUL" there by hand: char data[5120 * 4 * 2 + 1] in 13100, 18200,
19600, 19700, 32100 and 32200, and ct_data[16384 * 2 + 1] in 16900.
- The module hunks of 836f11de16 (13100, 18200) apply with an offset.
- caa6563ef9 also carries upstream's matching bound fixes for 00111,
01411, 01711, 05300, 05400, 11400 and 13600. I kept them, since they
are the same class of bug.
That is 23 files changed, 78 insertions and 33 deletions. None of the
patches in debian/patches touch src/modules/.

Testing: the patched 7.1.2 tree builds with "make" (gcc, amd64, sid)
with no new warnings in the touched modules. I have NOT run hashcat
with the patched modules (not even a parse round-trip of the example
hashes) and did not run any reproducer, so please treat the backport
as compile-tested only.

Regards

#1136004#19
Date:
2026-11-03 12:12:34 UTC
From:
To:
Hi,

A follow-up to my earlier mail, where I said the 7.1.2 backport was
compile-tested only. I have now run it.

Test: for each hash mode touched by the patch, I took hashcat's own
example hash (--example-hashes --machine-readable, field example_hash)
and ran

hashcat -m <mode> --left --potfile-path <empty file> <hashfile>

which parses the hash (module_hash_decode) and prints it back
(module_hash_encode). I ran it with the patched 7.1.2 build and with the
installed Debian hashcat 7.1.2+ds1-4, with an isolated HOME and an empty
potfile. The patched modules were built after the last patch was
applied.

Result: on both builds the output is identical to the input for all 23
modes:
Kerberos: 13100, 18200, 19600, 19700, 19800, 19900, 28800, 28900,
32100, 32200
PKZIP: 17200, 17210, 17220, 17225, 17230
others in the patch: 16900, 00111, 01411, 01711, 05300, 05400,
11400, 13600
The example hashes range from 46 to 1008 characters.

So the backport does not break parsing or re-encoding of valid hashes
for any mode it touches. I have not run a cracking session (there is no
OpenCL device here) and did not run any reproducer for the overflows
themselves.

Regards

#1136004#24
Date:
2026-11-15 05:49:10 UTC
From:
To:
We believe that the bug you reported is fixed in the latest version of
hashcat, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 1136004@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Daniel Echeverri <epsilon@debian.org> (supplier of updated hashcat package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@ftp-master.debian.org)
Format: 1.8
Date: Sun, 15 Nov 2026 00:03:37 -0500
Source: hashcat
Architecture: source
Version: 8.0.0+ds1-1
Distribution: unstable
Urgency: medium
Maintainer: Debian Security Tools <team+pkg-security@tracker.debian.org>
Changed-By: Daniel Echeverri <epsilon@debian.org>
Closes: 1136004 1145171
Changes:
 hashcat (8.0.0+ds1-1) unstable; urgency=medium
 .
   * New upstream version 8.0.0+ds1
     + Fixes CVE-2026-42482, CVE-2026-42483 and CVE-2026-42484
       (Closes: #1136004)
     + Fixes CVE-2026-68766, CVE-2026-68767 and CVE-2026-68768
       (Closes: #1145171)
   * debian/copyright:
     + Update Files-Excluded for 8.0.0
     + Update copyright entries for new and removed upstream files
     + Extend debian/* copyright years
   * debian/control:
     + Drop kfreebsd-any from Arch field
     + Remove redundant Priority: optional field
     + Bump Standards-Version to 4.7.4
     + Add ${perl:Depends} in Depends
   * debian/patches:
     + Remove fix_ftbfs_in_arm.diff and fix_correct_path_bridges.diff
     + Merge with upstream
     + Remove add_kfreebsd_support.diff, kFreeBSD is no longer supported
     + Refresh remaining patches
   * debian/rules:
     + Build with SHARED=0 to disable the shared library
     + Remove unnecessary files
   * debian/hashcat-data.install:
     + Install new upstream data files
   * debian/hashcat.install:
     + Install feeds and Python bridge files
   * debian/hashcat.links:
     + Add feeds link and remove old Python links
   * debian/hashcat.1:
     + Fix groff warning about escape character
Checksums-Sha1:
 1f3810ddd4d23dd812cac4becf54d58bcb266f45 2151 hashcat_8.0.0+ds1-1.dsc
 da9536d13596fd365d228ca2fb75acb0ae5df6a4 19787744 hashcat_8.0.0+ds1.orig.tar.xz
 30f5ada115834e2000ab9c42201dc4a493a63cc3 16548 hashcat_8.0.0+ds1-1.debian.tar.xz
 abd9c3eb7b92fc7a1c8cf62f2564c8a3bafed5e7 7742 hashcat_8.0.0+ds1-1_amd64.buildinfo
Checksums-Sha256:
 6c571d4bb533b46f1973dd0a7c6ea799c09c6cb75aa6b7892b04b739c507eeb0 2151 hashcat_8.0.0+ds1-1.dsc
 28bfb47edbdefdf5c282e0cdfea7c72eef57034be3feb0e1c5c10eb5907660b7 19787744 hashcat_8.0.0+ds1.orig.tar.xz
 0bf832af987b7eb4a812516bdcc11b30b534ac76f2bdef8b53e425d945bf8def 16548 hashcat_8.0.0+ds1-1.debian.tar.xz
 d241662a8f3da64fe0fc5645500e20d0b66682ec6455f8f1d1d72123ac78d768 7742 hashcat_8.0.0+ds1-1_amd64.buildinfo
Files:
 690335168bab947944ab3e9503cd1b89 2151 net optional hashcat_8.0.0+ds1-1.dsc
 e32cd161fb434d983c6d33284ec9ac26 19787744 net optional hashcat_8.0.0+ds1.orig.tar.xz
 47934b09abd8062499f382cc5063415f 16548 net optional hashcat_8.0.0+ds1-1.debian.tar.xz
 f1324e5ec53b262b48081e7a0dfd9925 7742 net optional hashcat_8.0.0+ds1-1_amd64.buildinfo
-----BEGIN PGP SIGNATURE-----
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=qHm+
-----END PGP SIGNATURE-----

#1136004#29
Date:
2026-11-15 06:22:01 UTC
From:
To:
Hi Daniel,

thanks for your update for hashcat in unstable. For the
security-tracker I have applied the tracking updates, in particular
CVE-2026-42482  was not present in v1.7.2 according to the upstream
information.  The other have identifiable commits, but I lack those
for CVE-2026-42483.

Can you please reference where this issue was fixed?

Regards,
Salvatore

#1136004#34
Date:
2026-11-15 15:54:55 UTC
From:
To:
Hello Salvatore!

El dom, 15 nov 2026 a la(s) 1:22 a.m., Salvatore Bonaccorso (
carnil@debian.org) escribió:

I did not find a clear reference for CVE-2026-42483 upstream, so I went
through the commits between 7.1.2 and 8.0.0, and I think these are the ones
that fix it:

adds the missing bounds check on account_info_len in modules 13100 and
18200[1]
fixes the out-of-bounds copy of the domain into a 128 byte stack buffer in
the etype 17/18 modules[2]

There is also a small follow-up for the first one, fixing an off-by-one[3]

Regards!

[1]:
https://github.com/hashcat/hashcat/commit/caa6563ef9dbd347f1c935212912cafd1f8a37cf
[2]:
https://github.com/hashcat/hashcat/commit/6a3994efda07d7200511de2f5382e724de6a85a4
[3]:
https://github.com/hashcat/hashcat/commit/836f11de165a1703d772fbafd82e17422b56f029

#1136004#39
Date:
2026-11-16 20:22:28 UTC
From:
To:
Hi Daniel,

Thank you!

Regards,
Salvatore