Home › Vulnerabilities

NervesHub (Nerves project) NervesHub CLI sign-in Flaw CVE-2026-119446: Update Now

CVE Critical Patch

The short version: CVE-2026-119446 is a critical-severity security flaw in NervesHub (Nerves project) NervesHub CLI sign-in. The National Vulnerability Database (NVD) has published it, and the score it lists is 9.4 out of 10. When FraudMeNot checked, CISA had not added CVE-2026-119446 to its Known Exploited Vulnerabilities catalog, so there was no public U.S. government confirmation of attacks yet. A fix is available, which is why this page focuses on who should update and how.

CVE-2026-119446 is a critical flaw in NervesHub, the open-source device management server from the Nerves project that organizations use to manage devices and deliver firmware to them through deployment groups. The bug is in the way NervesHub signs in its command-line tool. An outsider with no account could start a sign-in, send the resulting link to a real NervesHub user, and receive a working API token for that user's account as soon as the user clicked "Confirm login."

The NervesHub maintainers fixed the problem in version 2.5.3. The release notes describe it plainly: "CLI sign-in links gave an API token to whoever sent them."

How the sign-in flaw worked

The NervesHub command-line client, nh, signs in through the browser. According to the maintainers' advisory (GHSA-9vr9-463q-46qp), the client calls an API endpoint that "needs no authentication." The server answers with a polling token, a link and a six-digit confirmation code. The link contains the polling token.

Whoever opens that link while signed in to NervesHub sees the code and a "Confirm login" button. Clicking it makes NervesHub create an API token for the signed-in user and attach it to the pending login. The next time the client polls the server, it receives that token.

The weakness, in the advisory's words, is that "the token in the link is all the server needs to find and approve the login. The confirmation code is only displayed, and nothing compares it with anything." Because the attacker started the login, the attacker also has the code and can send it along with the link, so the code check on the page gives the victim false reassurance.

The attack still needs the victim's help. The confirmation form is protected against cross-site request forgery and the page cannot be embedded in another site, so, as the advisory notes, "the user has to open the link and click the button themselves." That makes it a phishing-style attack: a convincing message asking an operator to approve a sign-in.

Why the impact is serious

The CVE record describes the result as an attacker obtaining "an API token for another user's account." The CVSS score rates the impact as high for the NervesHub system itself and also for subsequent systems, which is how the scoring describes effects that reach beyond the server.

The maintainers also warn that cleaning up is not automatic: "Upgrading doesn't revoke tokens already handed out." A token obtained this way "is an ordinary API token, with a note the CLI chose, so an attacker's looks like any other."

Who is affected

The advisory and the CVE record list versions 2.5.0 through 2.5.2 of nerves_hub_web, including the published container image ghcr.io/nerves-hub/nerves-hub. Version 2.5.0 is the release that added CLI sign-in. Earlier versions do not have this particular flaw, although the same 2.5.3 release fixes other, separately numbered issues that reach further back.

Anyone operating a NervesHub server on an affected version should treat it as exposed, whether it serves one small team or a large device fleet.

Severity and who scored it

The CVE was assigned by the Erlang Ecosystem Foundation (EEF), the CVE Numbering Authority for the Erlang and Elixir ecosystem. EEF rated it CVSS 4.0 9.4 Critical, with the vector AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H. The "UI:A" part records that a user must actively take part, matching the click on "Confirm login." NVD lists the EEF score as secondary and had not added its own assessment when FraudMeNot checked. There is no CVSS 3.1 score for this CVE. The weakness is classified as CWE-384, session fixation.

The record credits Amr Kadry as finder and reporter, Josh Kalderimis as remediation developer, and Jonatan Männchen of the EEF as coordinator. CISA had not listed the CVE in its Known Exploited Vulnerabilities catalog, and neither the advisory nor the record mentions attacks.

What changed in 2.5.3

The fix, in commit 3131f17, changes the design rather than patching one check. Every CLI sign-in now uses the same address, /auth/cli, and the user must type the code the CLI shows. The polling token never reaches the browser. Codes are now eight digits from a cryptographic random source, and attempts are rate-limited. Before confirming, the page shows when and from which address the sign-in started, and warns when that address differs from the browser's.

The maintainers are candid about the limit: typing in a code "can't prove that the person typing started the login," which is a known limit of every device-code sign-in. The change removes the one-click link and gives users the context to notice a request they did not start.

Timeline

  • NervesHub 2.5.0 adds CLI sign-in.
  • 2.5.3 is released, the maintainers' advisory follows, and EEF publishes the CVE record.

What to do

These steps are for the people who operate a NervesHub server, and for every user of it.

  • Upgrade to 2.5.3 or later. Read the release's "Upgrading" section first, because it also changes sign-up settings, takes some devices off deployment groups that belong to other products, and makes the device API ignore fields it used to save.
  • If you cannot upgrade at once, block the entry point. The advisory's workaround is to block POST /api/auth/cli_session at your reverse proxy. It is the only request that starts a CLI sign-in, so no new links can be created while it is blocked. CLI users can still sign in with email and password (nh user auth) or with an API token created on the account page.
  • Review existing tokens. Because upgrading does not revoke anything, ask every user to open the account page, look through their API tokens and delete any they do not recognise.
  • Warn users about approval requests. Tell operators never to confirm a CLI sign-in, or enter a code, that someone else sent them, even after upgrading.

What the official record says about CVE-2026-119446

The record was filed by 6b3ad84c-e1a6-4bf7-a703-f496b71e49db. In its own words: "Session Fixation vulnerability in the CLI login flow in nerves-hub nerves_hub_web allows an unauthenticated attacker to obtain an API token for another user's account via a crafted login link. Starting a CLI login with POST /api/auth/cli_session needs no authentication and returns a polling token, a confirmation URL that carries that token, and a confirmation code. When a signed-in user opens that URL and clicks Confirm, NervesHub.Accounts.verify_cli_session_token/2 creates an API token for the user's account and stores it on the login, and the next poll of GET /api/auth/cli_session/<token> returns it to whoever started the login. The confirmation code is only displayed and is never compared with anything, so it cannot show that the person confirming is the one who started the login. The confirm form carries a CSRF token and the page cannot be framed, so the attacker has to persuade the user to open the link and click the button themselves. This issue affects nerves_hub: from 2.5.0 before 2.5.3."

Why NervesHub CLI sign-in admins should care

Severity scores use the Common Vulnerability Scoring System (CVSS). 6b3ad84c-e1a6-4bf7-a703-f496b71e49db rated CVE-2026-119446 9.4 (Critical) under CVSS 4.0. Reading that rating in everyday terms: it can be reached over a network, including the internet; an attacker does not need an account; a user has to take a deliberate action. If it is abused, the rated damage to keeping data private is high, to keeping data unaltered is high, and to keeping NervesHub CLI sign-in running is high. Scoring string: CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X.

Affected NervesHub CLI sign-in releases

According to the record, these releases are vulnerable:

  • 2.5.0 up to, but not including, 2.5.3
  • 2.5.0 up to, but not including, 2.5.3
  • e6bf11a812d7aa8b69d854ba8ef02d65516f5564 up to, but not including, 3131f1701194b581ab4b3e5e9d81194fc9c83508

The type of bug behind CVE-2026-119446

NVD classifies it using MITRE's Common Weakness Enumeration:

  • CWE-384, Session Fixation. MITRE describes this class of bug as follows: "Authenticating a user, or otherwise establishing a new user session, without invalidating any existing session identifier gives an attacker the opportunity to steal authenticated sessions."

Vendor remediation for CVE-2026-119446

The vendor advisory remediation for operators of NervesHub CLI sign-in, quoted for this page, is: "Patch: upgrade nerves_hub_web / ghcr.io/nerves-hub/nerves-hub from 2.5.0-2.5.2 to 2.5.3 or later (commit 3131f17). Read the 2.5.3 Upgrading notes first. Workaround (maintainer advisory): block POST /api/auth/cli_session at the reverse proxy until upgrade; CLIs can use nh user auth or an account-page API token. After upgrading, users must review and delete unrecognised API tokens (upgrade does not revoke them). Who acts: NervesHub server operators and all users. Score: EEF (CNA) CVSS 4.0 9.4 Critical, UI:A (NVD Secondary). Not in CISA KEV when checked."

What to do about it

  • Update now: Patch: upgrade nerves_hub_web / ghcr.io/nerves-hub/nerves-hub from 2.5.0-2.5.2 to 2.5.3 or later (commit 3131f17). Read the 2.5.3 Upgrading notes first. Workaround (maintainer advisory): block POST /api/auth/cli_session at the reverse proxy until upgrade; CLIs can use nh user auth or an account-page API token. After upgrading, users must review and delete unrecognised API tokens (upgrade does not revoke them). Who acts: NervesHub server operators and all users. Score: EEF (CNA) CVSS 4.0 9.4 Critical, UI:A (NVD Secondary). Not in CISA KEV when checked.
  • Who should act: IT teams, managed service providers, and developers who run or ship NervesHub CLI sign-in. Check every place it runs, including servers, build systems, containers, and older machines nobody watches closely.
  • Read the official advisory from NervesHub (Nerves project) (linked under Sources) for exact fixed builds and any temporary workaround if you cannot patch today.
  • After patching, confirm the new version is actually running and review logs from the patch release onward for anything unusual.
  • Home users: you do not need to buy or download anything because of CVE-2026-119446. A caller, text, or pop-up claiming your computer is "infected" with this CVE and asking for money or remote access is a scam.

CVE-2026-119446 record details from CISA and NVD

  • Attack complexity for CVE-2026-119446: low
  • CVE-2026-119446 attack requirements: none
  • Impact on connected systems: secrecy high, integrity high, availability high.
  • Who mapped the weakness type: 6b3ad84c-e1a6-4bf7-a703-f496b71e49db (Secondary)
  • References in the NVD record: 5.

What the official record shows

  • NVD has published CVE-2026-119446.
  • When FraudMeNot checked the CISA KEV catalog, CVE-2026-119446 was not listed.
  • The NVD record status is Received.

Official references

Links listed in the NVD record (labels in brackets come from NVD):

Sources

General education only, not security advice. Facts checked against the sources above.