/

CVE-2026-49973

Severity:

CRITICAL RISK

(9.4)

Missing Authentication (CWE-306) in Hermes WebUI Setup

CVE-2026-49973

Deferred

Vulnerability type

Category:

Missing Authentication

CWE:

CWE-306

Impact:

Authentication Bypass and Account Takeover

Severity:

CRITICAL RISK

(9.4)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

Hermes WebUI contains an improper access control flaw in its initial setup flow that allows unauthenticated attackers to set the first administrative password via the settings API.

By sending a crafted request during the first-run setup window, an attacker can take over the instance, establish a valid session, and prevent the legitimate operator from completing setup.

Affected context

The vulnerability arises in Hermes WebUI deployments that are in their initial "first run" setup state and expose the web interface over a reachable network.

During this setup window, the settings API endpoint accepts the _set_password parameter without enforcing authentication or network origin restrictions, allowing anyone with network access to the interface to attempt to complete the setup.

Exploit preconditions

  • Attack surface: HTTP(S) settings API endpoint of Hermes WebUI during first-run setup
  • Required access: Network reachability to the WebUI (local network, VPN, or internet-exposed)
  • Trigger point: Submitting a POST request with the _set_password parameter to the settings endpoint while no administrator account is yet finalized
  • Why exploitation is possible: The initial setup flow does not enforce authentication or origin checks before accepting a password hash and issuing a session, allowing unauthenticated callers to claim the first administrator credentials

Why this happens

This issue stems from missing or incomplete authentication checks on a security-critical action—setting the initial administrative password.

Instead of requiring proof of control (such as a bootstrap token, local-only access, or prior authentication), the application trusts any caller to submit _set_password during the first-run configuration phase.

This maps to CWE-306: Missing Authentication for Critical Function, where a function that establishes or controls authentication state is exposed without proper access control or origin validation.

Potential impact

  • Unauthenticated attackers may set an arbitrary administrative password hash.
  • Attackers can obtain a valid session cookie and full administrative control of the instance.
  • Legitimate operators can be locked out from completing or accessing their own deployment.
  • Compromised instances may be used to exfiltrate or alter data managed through the WebUI.
  • The compromised interface could become a foothold for broader infrastructure attacks depending on integrations.

General mitigation guidance

Mitigation generally focuses on hardening the initial setup process and reducing trust in unauthenticated network traffic:

  • Require an out-of-band bootstrap secret, setup token, or one-time code before allowing password initialization.
  • Restrict the first-run setup interface to trusted origins or localhost-only access by default.
  • Enforce authentication or strong capability checks before accepting any password-setting or account-creation actions.
  • Separate configuration write operations behind explicit privilege checks, even during setup.
  • Apply least-privilege principles to internal components so that compromising the WebUI does not grant unnecessary downstream access.

How teams detect issues like this

Detection can be approached through a mix of code-level and operational controls:

  • Static analysis and code review rules that flag routes or endpoints performing authentication or password-related operations without explicit access control middleware.
  • Security reviews focused on "first run" or installation flows, looking for unauthenticated configuration or account-creation steps.
  • Web server and application logs that highlight early POST requests to the settings endpoint, especially those containing _set_password from unexpected IP ranges.
  • Automated security testing (DAST) designed to probe installation and onboarding paths for missing authentication and origin checks.

Sources

SourceLink
github.comgithub.com/nesquena/hermes-webui/commit/1126e541325d401538f6a272a9c024c37d47ae08
github.comgithub.com/nesquena/hermes-webui/pull/3964
github.comgithub.com/nesquena/hermes-webui/pull/3973
github.comgithub.com/nesquena/hermes-webui/releases/tag/v0.51.358
vulncheck.comvulncheck.com/advisories/hermes-webui-unauthenticated-password-takeover-via-api-settings

Frequently asked

questions

Yes, if the Hermes WebUI setup interface is reachable over a network, an attacker can exploit the issue remotely during the first-run setup window.

The attacker only needs network access to the WebUI endpoint—no prior credentials or local system access is required—to attempt setting the initial password.

The safest pattern is to require a protected bootstrap mechanism (such as a setup token, local-only access, or pre-configured secret) before any password-setting or administrator-creation action can succeed, and to treat these flows like other privileged operations with full authentication and authorization checks.

Design installation and onboarding flows with the same rigor as normal authentication paths: mandate explicit access control for all configuration-changing endpoints, restrict default exposure of setup interfaces, and include targeted security reviews and automated tests for "first run" functionality.

Find out what's already
exploitable in your codebase.