/

CVE-2026-35579

Severity:

CRITICAL RISK

(9.8)

Broken / Improper Authentication in CoreDNS DoH Transports

CVE-2026-35579

Modified

Vulnerability type

Category:

Broken / Improper Authentication

CWE:

CWE-287, CWE-303

Impact:

Authentication Bypass Leading to Unauthorized DNS Zone Transfers and Updates

Severity:

CRITICAL RISK

(9.8)

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

CoreDNS versions prior to 1.14.3 contain an authentication flaw in the handling of TSIG records over gRPC, QUIC, DoH, and DoH3 transports, where message authentication codes are never correctly verified.

A remote attacker can spoof TSIG-authenticated DNS requests and bypass protections around zone transfers, dynamic updates, or other TSIG-gated behavior, enabling unauthorized DNS changes and data exfiltration.

Affected context

This vulnerability affects CoreDNS deployments that:

  • Enable gRPC, QUIC, DoH, or DoH3 listeners, and
  • Rely on TSIG to protect sensitive operations such as AXFR/IXFR, dynamic DNS updates, or plugin logic that checks TSIG authentication.

The issue manifests in environments where TSIG is assumed to provide strong message authentication, but the server-side transport code validates only the TSIG key name (or does not inspect TSIG at all) instead of verifying the HMAC. Deployments that do not use TSIG, or do not expose these transports, are less exposed to this specific flaw.

Exploit preconditions

  • Attack surface: Network-accessible CoreDNS endpoints using gRPC, QUIC, DoH, or DoH3.
  • Required access: Ability to send crafted DNS queries over the affected transport to the CoreDNS instance (no valid TSIG key material required; for DoH/DoH3 not even a valid key name is needed).
  • Trigger point:
  • For gRPC and QUIC: Send DNS messages with a TSIG record whose key name matches any configured TSIG key; the server does not verify the HMAC.
  • For DoH and DoH3: Send any DNS message containing a TSIG record; the TSIG is never inspected and is treated as valid.
  • Why exploitation is possible: The implementation treats TSIG as authenticated based on key name presence (gRPC/QUIC) or unconditionally (DoH/DoH3), never invoking proper TSIG MAC verification (dns.TsigVerify). This leads to a logic-level authentication bypass for TSIG-gated functionality.

Why this happens

The bug stems from incomplete and incorrect implementation of TSIG authentication logic within CoreDNS transport layers.

For gRPC and QUIC, the code only checks whether a TSIG key name exists in configuration but omits the call to the TSIG verification routine (e.g., dns.TsigVerify()), leaving tsigStatus in a state interpreted as "success" regardless of the MAC. For DoH and DoH3, the DoHWriter.TsigStatus() method simply returns nil, indicating success without evaluating the TSIG record at all.

This behavior maps to CWE-287: Improper Authentication, where the system bases authentication decisions on incomplete checks (key name presence or unconditional success) instead of verifying cryptographic proof of authenticity. Conceptually, it is a logic flaw in the authentication pipeline: an authentication token (TSIG) is accepted without validating its integrity and authenticity.

Potential impact

  • Unauthorized AXFR/IXFR zone transfers that may expose internal DNS data.
  • Unauthorized dynamic DNS updates that can poison DNS records or redirect traffic.
  • Bypass of TSIG-protected plugin logic, potentially enabling unintended operations or policy violations.
  • Facilitation of phishing, man-in-the-middle, or internal service impersonation via malicious DNS entries.
  • Cascading compromise in environments where other security controls assume the integrity of DNS data.

General mitigation guidance

Effective mitigation focuses on restoring strong, mandatory authentication semantics for TSIG-protected operations and narrowing the exposure of high-privilege transports:

  • Prefer transport paths and protocol handlers that strictly validate TSIG MACs and enforce clear authentication outcomes for privileged DNS operations.
  • For environments requiring TSIG protection, avoid routing sensitive operations (like zone transfers or dynamic updates) over transports that do not robustly enforce TSIG verification.
  • Apply network segmentation and least-privilege design for DNS: expose high-privilege endpoints only on restricted networks and to trusted peers.
  • Use defense-in-depth for DNS update and zone-transfer workflows, such as additional authorization checks, policy-based access control, or out-of-band approvals where feasible.
  • Treat TSIG keys and configuration as high-value secrets and ensure that policies do not implicitly trust requests solely based on protocol or transport.

How teams detect issues like this

Detection typically combines configuration review, traffic inspection, and code-level analysis:

  • Configuration and architecture review:
  • Identify CoreDNS instances that enable gRPC, QUIC, DoH, or DoH3.
  • Map where TSIG is expected to enforce access control for zone transfers, updates, or plugins, and check if these controls rely on the affected transports.
  • Static analysis and code review:
  • Search the codebase for TSIG handling paths where authentication decisions are made without explicit MAC verification calls.
  • Flag patterns where TSIG-related status functions always return success or where TSIG records are parsed but not validated.
  • Runtime and traffic monitoring:
  • Log and analyze zone-transfer and dynamic-update events, especially those originating from unexpected sources or over DoH/DoH3.
  • Use DNS-aware monitoring tools or proxies to verify that TSIG-protected operations only occur when valid TSIG signatures are present.
  • Automated security checks:
  • Integrate SAST or linting rules that detect incomplete cryptographic or authentication checks.
  • Use integration tests that attempt TSIG-protected operations with invalid or missing MACs to ensure they fail as expected.

Sources

SourceLink
github.comgithub.com/coredns/coredns/security/advisories/GHSA-vp29-5652-4fw9
access.redhat.comaccess.redhat.com/errata/RHSA-2026:25127
access.redhat.comaccess.redhat.com/security/cve/CVE-2026-35579
bugzilla.redhat.combugzilla.redhat.com/show_bug.cgi?id=2466905
security.access.redhat.comsecurity.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-35579.json

Frequently asked

questions

Yes, if the CoreDNS instance exposes gRPC, QUIC, DoH, or DoH3 endpoints to untrusted networks. An attacker who can reach these ports can attempt to bypass TSIG protections without possessing valid TSIG keys.

No. For gRPC and QUIC, the attacker only needs to know a TSIG key name configured on the server, not the key material. For DoH and DoH3, the bar is even lower: any TSIG-bearing request may be accepted as authenticated regardless of key name or MAC.

Ensure that all authentication decisions are tied to explicit, mandatory verification of cryptographic proofs—here, invoking TSIG MAC verification and basing access control solely on its result. Authentication status should never default to success and should fail closed if verification cannot be completed.

Treat authentication and cryptographic verification as first-class, testable components: centralize auth logic, enforce fail-closed semantics, include negative tests (invalid signatures, wrong keys), and integrate static/dynamic checks for code paths that bypass verification routines. Additionally, adopt code review practices and tooling that flag authentication flows lacking explicit verification or relying only on identifiers like key names.

Find out what's already
exploitable in your codebase.