Raw vector
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:HCVSS and EPSS are reproduced from their sources (NVD, FIRST EPSS). EPSS, Forum of Incident Response and Security Teams (FIRST), epss.first.org. Risk Priority is our own derived reading, not an NVD score.
Summary
CVE-2023-27584 is a critical-severity Use of Hard-coded Cryptographic Key (CWE-321) vulnerability in Linuxfoundation Dragonfly. Its CVSS base score is 9.8 (Critical).
Operationally, exploitation aligns with the MITRE ATT&CK technique Remote Services (T1021); ranked in the top 2% of CVEs by exploit likelihood; it is not currently listed in the CISA KEV catalog; a public proof-of-concept is referenced.
The strongest mitigations our analysis identified map to IA-5 (Authenticator Management) and SC-12 (Cryptographic Key Establishment and Management) — see the control section below for these in your framework.
Deeper analysis AI-assisted summary
Synthesised by an AI model from the NVD description and linked references — a reading aid, not an authoritative source.
Dragonfly, an open source P2P-based file distribution and image acceleration system hosted by the CNCF, contains a hardcoded JWT secret key value of "Secret Key" used for user authentication. This flaw, tracked as CVE-2023-27584 and assigned CVSS 9.8, corresponds to CWE-321 and CWE-798 and permits complete authentication bypass in all versions prior to the fix.
An unauthenticated remote attacker can exploit the static key to forge valid JWT tokens and perform any action with administrative privileges, including arbitrary operations on the Dragonfly deployment. The vulnerability requires no user interaction or special network positioning beyond network reachability to the affected service.
The official security advisory and release notes for Dragonfly2 version 2.0.9 state that the issue has been resolved in that version and that no workarounds exist, directing all users to upgrade immediately. The associated EPSS score has reached 0.6618 without a documented low-to-high trajectory after disclosure.
OWASP Top 10 for Web (2025)
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2024-2812
Vulnerability Data
Dragonfly is an open source P2P-based file distribution and image acceleration system. It is hosted by the Cloud Native Computing Foundation (CNCF) as an Incubating Level Project. Dragonfly uses JWT to verify user. However, the secret key for JWT, "Secret…
more
Key", is hard coded, which leads to authentication bypass. An attacker can perform any action as a user with admin privileges. This issue has been addressed in release version 2.0.9. All users are advised to upgrade. There are no known workarounds for this vulnerability.
- CWE(s)
Related Threats
MITRE ATT&CK Enterprise Techniques
MITRE ATT&CK v20.0
CVEs Like This One
Affected Assets
Mitigating Controls
Mitigating Controls (NIST 800-53 r5)
Our mappings, made for this CVE.
Requiring cryptographic keys to be established and managed according to defined requirements prevents developers from embedding static unchangeable keys.
Authenticator management requires secure distribution and handling of credentials, structurally discouraging hard-coded values.
Mitigating Controls (NIST CSF 2.0)
Our mappings, derived directly from the weakness types (CWEs) cited in the NVD entry via our CWE→CSF cross-walk — links open the control.
Secure-SDLC activities such as code review and secret scanning directly prevent embedding static keys.
PR.AA-01's credential/key-management processes can reduce the incentive to embed secrets but do not address or detect hard-coded values in source code, so the weakness remains fully possible.
Data-at-rest protection policies require proper key management and therefore discourage hard-coded keys.
Data-in-transit protection similarly depends on non-hard-coded keys for encryption.
Configuration baselines and reviews can prohibit hard-coded keys in deployed artifacts.
Mitigating Controls (ISO/IEC 27001:2022 Annex A)
Our mappings, derived directly from the weakness types (CWEs) cited in the NVD entry via our CWE→ISO cross-walk — links open the control.
A.8.24's key management section explicitly mandates proper key generation, storage, distribution, and lifecycle processes — the antithesis of hard-coding keys into products — so a well-implemented cryptography policy prevents hard-coded keys in anything the organization designs or builds; the bounded remainder is third-party, COTS, and legacy components that arrive with keys already embedded, mirroring the CWE-328 anchor.
A.8.29's SOURCE + RUNTIME reach finds hard-coded cryptographic keys in first-party source via SAST, code review, and penetration testing — the same mechanism that graded A.8.29 finds CWE-798 = mostly; the bounded remainder is keys assembled at build time, baked into images as configuration, or shipped inside binary-only third-party components.
Education on secure configuration practices discourages technical staff from embedding or relying on hard-coded credentials in systems and applications.
A.8.28 explicitly prohibits hard-coded passwords as an example of insecure design techniques and mandates secure coding standards that would cover hard-coded cryptographic keys, but as a practice-based control it is fallible and depends on developer compliance, with the strongest detection coming from A.8.29 security testing.
A.8.30 carries its effect through the secure-coding practices (A.8.28) its contract requires of the supplier, which reduces hard-coded credentials in delivered code the way those practices do; the clause depends on the supplier actually implementing secure coding, and secure coding is a fallible practice, so a minority of instances are prevented.
A.8.8's reach is DISCLOSURE: it finds hard-coded keys only where a vendor has published a CVE/advisory for a third-party component in the operator's inventory. CWE-321 does surface as CVEs (multiple observed examples), so the disclosed third-party share is real, but it is a minority of all instances — first-party code is outside A.8.8's reach (that is A.8.28/A.8.29), and many hard-coded keys remain unpublished. `full` is out of reach on disclosure-only scanning.