MacCMS v10: anyone can reset member passwords without logging in
A flaw in the MacCMS website software lets an outsider take over any member's account. The attacker only needs the victim's email or phone number, not a password.
| Severity | CriticalCVSS 9.3 of 10 |
|---|---|
| Being exploited? | UnclearNo reports either way yet |
| Fix available? | YesMacCMS v10 2026.1000.4054 or later · since 26 Oct 2026 |
| How sure is this? | Official sourceBacked by an official source |
How the attack works simplified
- 1Attacker requests codeSends recovery request to their own mailbox
- 2Code sent to attackerRecovery code goes to attacker-controlled address
- 3Code submitted for victimUses victim's email or phone in reset request
- 4Check skips recipient matchmsg_to is never compared, so code is accepted
- 5Password resetAttacker sets new password for victim account
- ResultAttacker controls the victim's account
What happened
In MacCMS v10 before 2026.1000.4054, check_msg() never compares msg_to, so a recovery code sent to the attacker's own mailbox via user/findpass_msg is accepted by user/findpass_reset for a victim's email or phone, enabling an authentication bypass.
Are you affected?
You are affected if you run MacCMS v10 earlier than 2026.1000.4054 with member/password-recovery features enabled.
MacCMS v10, all versions before 2026.1000.4054
What to do
- Update MacCMS v10 to 2026.1000.4054 or later
- Review member accounts and password reset activity for unexpected changes
- Force password resets for accounts that may have been targeted
Timeline
- 26 Oct 2026, 14:16 UTCPublicly disclosed
- 26 Oct 2026, 14:16 UTCReported by NVD · NVD
- 26 Oct 2026, 14:21 UTCPicked up by Zero Radar
Sources 1 report, oldest first
Official vendor, government CERT or advisory database
- CVE-2026-116214NVD · 26 Oct 2026, 14:16 UTC
Technical details
- ID
- CVE-2026-116214
- CVSS
- 9.3 / 10
- EPSS
- 0.5% chance of exploitation within 30 days
- Disclosed
- 26 Oct 2026, 14:16 UTC
- On Zero Radar
- 26 Oct 2026, 14:21 UTC (4 min after disclosure)
- Last change
- 26 Oct 2026, 14:21 UTC
The plain-language parts and the attack diagram were written by an AI model (anthropic/claude-haiku-5.5) from the sources above. Facts like versions, exploitation and dates come from the sources themselves. How Zero Radar works