Security Advisory

CVE-2020-10134: Bluetooth Method Confusion MITM

Last updated:
  • CVE-2020-10134

CVE-2020-10134 is a flaw in the Bluetooth Core specification v5.2 and earlier: an attacker in radio range during first-time pairing can make one device use Numeric Comparison and the other Passkey Entry, and if the user types the displayed number as the passkey, the attacker sits in the middle of both connections. NVD scores it 6.3. There is no code patch; the defence is clear pairing UI and out-of-band pairing.

Vulnerability summary

CVSS base score

6.3

Medium severity

EPSS score (FIRST)

0.7%

Probability of exploitation in the wild in the next 30 days

CISA KEV catalog

No

Not listed

Published

Vulnerability ID: CVE-2020-10134 · Source: NVD

Affected components

ComponentCPEVersion range
Bluetooth Core Specification (BR/EDR and LE pairing)cpe:2.3:a:bluetooth:bluetooth_core:*:*:*:*:le:*:*:*Up to and including 5.2 for the BR, EDR and LE configurations (NVD CPE); LE Secure Connections 4.0 to 5.2 and BR/EDR Secure Simple Pairing 2.1 to 5.2 per Red Hat bug 1832216
BlueZ (Linux Bluetooth stack) in Ubuntupkg:deb/ubuntu/bluezMarked 'Vulnerable, fix deferred' on Ubuntu 16.04, 18.04, 20.04, 22.04, 24.04 and 26.04 LTS (Ubuntu CVE tracker, as of 2026-11-20); Debian lists no affected package ('Bluetooth protocol issue')
Bluetooth chips and embedded stacks (host and controller)—Any implementation offering Numeric Comparison or Passkey Entry pairing; CERT/CC VU#534195 lists Broadcom, Cypress, Espressif, Intel, MediaTek, Microchip, Qualcomm and Zephyr Project with status Unknown (as of 2026-11-20)

What does this mean for affected devices?

CVE-2020-10134 is a design flaw in Bluetooth pairing, not a bug in one product. The CVE record says pairing "in Bluetooth® Core v5.2 and earlier may permit an unauthenticated attacker to acquire credentials with two pairing devices via adjacent access" when the attacker starts a different pairing method with each peer and the user "erroneously completes both pairing procedures with the MITM using the confirmation number of one peer as the passkey of the other." The record was published on 2020-05-19 (NVD, CVE record).

How does the method confusion attack work?

CERT/CC VU#534195 explains that two association models, Numeric Comparison (NC) and Passkey Entry (PE), are affected in both LE Secure Connections and BR/EDR Secure Simple Pairing. The attacker sits between two devices that are pairing for the first time, with no stored long term key or link key, negotiates NC with one and PE with the other, and relies on the user typing the six-digit NC value shown on one device as the passkey on the other. CERT/CC gives the root cause: NC and PE "use the same form of check value, the model used is not indicated to the user", and neither device authenticates which model the peer used. One device must accept passkey entry and the other must show six decimal digits; in BR/EDR, only a device acting as a keyboard for pairing can enter the passkey. Red Hat's tracking bug quotes the report as covering LE Secure Connections in Core 4.0 through 5.2 and BR/EDR Secure Simple Pairing in Core 2.1 through 5.2 (Red Hat bug 1832216). The researchers, Ludwig Peuckert and Maximilian von Tschirschnitz, published a proof of concept, BThack, built on a modified BTstack.

How severe is CVE-2020-10134?

NVD's Primary score is CVSS v3.1 6.3 (CVSS:3.1/AV:A/AC:L/PR:N/UI:R/S:U/C:H/I:L/A:N), the same vector that CERT/CC, the CNA, assigned; v2 is 4.3. NVD classes the weakness as CWE-436 (interpretation conflict), while CERT/CC uses CWE-351 (insufficient type distinction). The attacker must be in radio range during first-time pairing and the user has to make the mistake, which keeps the score medium. If it works, CERT/CC says the attacker "could be able to initiate any Bluetooth operation on either attacked device exposed by the enabled Bluetooth profiles."

Which firmware and devices are affected?

NVD's CPE covers the Bluetooth Core specification BR, EDR and LE configurations up to and including 5.2. Because the flaw is in the specification, any product that offers NC or PE pairing can be exposed, typically a device with a six-digit display paired against one with a keypad. CERT/CC lists the Bluetooth SIG as Affected; seven vendors replied Not Affected (CoreOS, Illumos, Joyent, LANCOM Systems, Pulse Secure, Sierra Wireless, Zyxel), and the Bluetooth chip and stack suppliers it notified, including Broadcom, Cypress Semiconductor, Espressif Systems, Intel, MediaTek, Microchip Technology, QUALCOMM, Android Open Source Project and the Zephyr Project, remain Unknown as of 2026-11-20. No device vendor or CISA advisory names products for this CVE as of 2026-11-20. Distribution trackers show no code fix: the Debian security tracker lists no package and notes "Bluetooth protocol issue", Ubuntu marks bluez "Vulnerable, fix deferred" on its supported LTS releases with the note that fixes "may be strictly user-interface messages", and Red Hat closed its bug as WONTFIX.

How do you fix CVE-2020-10134?

There is no firmware patch that removes it. CERT/CC tells host and controller suppliers to follow the Bluetooth SIG statement and downstream vendors to ask their suppliers for updates. Red Hat's mitigation is out-of-band (OOB) pairing where possible, or turning Bluetooth off where it is not needed. The researchers' BThack notes say vendors can only educate users and "visualize the utilized pairing method prominently", which lets an attentive user notice the attack; for a device design that means a displayed number is labelled as a value to compare, never one to type. Our NIST SP 800-121 Bluetooth security guide lists this notice with the other SIG specification flaws, and the BLE security guide for IoT devices covers pairing choices. Related specification flaws on this site are the LE passkey confusion CVE-2022-25836, passkey entry impersonation CVE-2020-26558 and BIAS, CVE-2020-10135.

Is CVE-2020-10134 exploited?

It is not in the CISA KEV catalog as of 2026-11-20 (catalog version 2026.11.18, 1,786 entries). FIRST EPSS gives it 0.0067, the 52.8th percentile, in the 2026-11-19 dataset (CISA KEV, FIRST EPSS). A public proof of concept exists in the researchers' BThack repository.

How to fix this class of issueNo dedicated SoftWSP guide covers this weakness class; the fix is usually the vendor patch or firmware update referenced above. Our related guide explains how to plan and verify it: NIST SP 800-40 Rev. 4: enterprise patch management planning, explained.

Evidence

  • NVD (read 2026-11-20): status Modified; published 2020-05-19; NVD Primary CVSS v3.1 6.3 CVSS:3.1/AV:A/AC:L/PR:N/UI:R/S:U/C:H/I:L/A:N, v2 4.3; CERT/CC ([email protected]) Secondary 6.3 same vector; CWE NVD CWE-436, CERT/CC CWE-351; CPE bluetooth_core br/edr/le up to and including 5.2; references CERT/CC VU#534195 and the Bluetooth SIG statement. — NVD
  • CERT/CC VU#534195 (read 2026-11-20; released 2020-05-18, revised 2020-05-26): NC and PE affected in LE Secure Connections and BR/EDR Secure Simple Pairing; attacker negotiates NC with one device and PE with the other; root cause same check value, model not shown to user, peer model not authenticated; Bluetooth SIG Affected; 7 vendors Not Affected (CoreOS, Illumos, Joyent, LANCOM, Pulse Secure, Sierra Wireless, Zyxel); 211 vendors Unknown including Broadcom, Cypress, Espressif, Intel, MediaTek, Microchip, QUALCOMM, AOSP, Zephyr Project; reported by Ludwig Peuckert and Maximilian von Tschirschnitz. — CERT/CC
  • Red Hat bug 1832216 (read 2026-11-20): CLOSED WONTFIX (2021-10-28); LE Secure Connections in Core 4.0 through 5.2 and BR/EDR Secure Simple Pairing in Core 2.1 through 5.2; mitigation: use OOB pairing if possible, or disable Bluetooth. — Red Hat
  • Debian security tracker (read 2026-11-20): no source package listed; note 'Bluetooth protocol issue'. Ubuntu CVE tracker (read 2026-11-20): bluez 'Vulnerable, fix deferred' on 16.04, 18.04, 20.04, 22.04, 24.04 and 26.04 LTS; note that fixes may be strictly user-interface messages; no software fix available as of 2021-12-08. — Debian / Ubuntu
  • BThack README (read 2026-11-20): proof of concept of the Method Confusion attack using a customized BTstack; states no fix available that would not massively affect backwards compatibility, and that vendors can only educate users and visualize the utilized pairing method prominently. — BThack (researchers)
  • CVE record (checked 2026-11-20): CNA CERT/CC, published 2020-05-19, CNA CVSS v3.1 6.3 and CWE-351, description 'Pairing in Bluetooth® Core v5.2 and earlier ...'. — CVE.org
  • Not in the CISA KEV catalog, version 2026.11.18 (1,786 entries), checked 2026-11-20. — CISA KEV
  • EPSS 0.0067, percentile 0.52791, dataset 2026-11-19, read 2026-11-20. — FIRST EPSS

Related advisories

Sources

  1. NVD
  2. CERT/CC Vulnerability Note VU#534195: Bluetooth method confusion attacks
  3. Bluetooth SIG statement: Method Confusion Pairing Vulnerability
  4. Red Hat bug 1832216: CVE-2020-10134
  5. Debian security tracker: CVE-2020-10134
  6. Ubuntu CVE tracker: CVE-2020-10134
  7. BThack: proof of concept of the Method Confusion attack
  8. CVE.org record CVE-2020-10134 (CNA: CERT/CC)
  9. CISA KEV
  10. FIRST EPSS

Frequently asked questions

Is there a firmware update that fixes CVE-2020-10134?

No code fix removes it, because the flaw is in the Bluetooth specification. Red Hat closed its bug as WONTFIX, Ubuntu marks bluez fix deferred and Debian lists it as a Bluetooth protocol issue (as of 2026-11-20). Mitigation is pairing UI design and out-of-band pairing.

Which devices can be attacked?

Two devices pairing for the first time with Numeric Comparison or Passkey Entry, where one can enter a passkey and the other shows six digits. Devices already bonded, and pairings using out-of-band data, are outside the attack described by CERT/CC.

What should a device maker change in the pairing UI?

Show which pairing method is in use, as the researchers recommend in their BThack notes: label a displayed number as a value to compare, never to type, and label a passkey prompt as entry of a code shown by the peer. Red Hat adds out-of-band pairing where the product allows it.

Did chip vendors confirm whether their stacks are affected?

Not in CERT/CC VU#534195: Broadcom, Cypress, Espressif, Intel, MediaTek, Microchip, Qualcomm and the Zephyr Project are listed with status Unknown as of 2026-11-20. The Bluetooth SIG itself is listed as Affected, since the flaw is in the specification.