Threats / JFrog / CVE-2026-66384

CVE-2026-66384 · EUVD no mirror located · GCVE no mirror located Verified 2026-08-28

JFrog Artifactory vulnerability

JFrog Artifactory contains an improper limitation of a pathname to a restricted directory vulnerability. This can allow an authenticated user to write data outside the intended Docker cache path under specific remote-repository conditions.

Verdict

Today item — known-exploited.

Confirmed exploited in the wild (CISA KEV). Prioritize remediation now and verify exposure.

CISA KEV Yes · 2026-08-27⁵EPSS 0.00543 (verify live)⁶CVSS 5.3³CISA SSVC Exploitation: active · Automatable: no · Impact: partial⁴

Is it exploitable?

— the evidence, ranked above the score

Reported exploitation

2 independent public reports of in-the-wild exploitation are cataloged.Distinct reporting sources (vendor, incident response, government); open them for the underlying claims.

previdian.com ↗cisa.gov ↗Confirmed

Exploited in the wild

Listed in the CISA Known Exploited Vulnerabilities catalog (added 2026-08-27).

CISA KEV ↗Confirmed

Probability (EPSS)

EPSS 0.00543 — modeled likelihood of exploitation activity.EPSS is a daily-changing model output — open the source for today's value.

FIRST EPSS ↗High

Severity / affected

CVSS 5.3 (CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:H/A:N) · Affected: JFrog, Artifactory. Confirm exact fixed builds in the vendor advisory.CVSS — CNA-assigned (CVE.org), version 3.1, set at disclosure.

CVE.org ↗Reported

Weakness (CWE)

Mapped to CWE-22 Path Traversal — weakness family: Path traversal / file.CWE — CNA-assigned (CVE.org); the weakness class drives how the flaw is exploited.

CVE.org ↗Reported

CISA SSVC (triage disposition)

Exploitation: active · Automatable: no · Technical Impact: partialCISA's own stakeholder-specific triage inputs, published in the CVE record (CISA Coordinator role). NOTE: this is CISA's three decision points; the final Act/Attend/Track decision also requires a Mission & Well-being input not published here, so no overall decision is asserted.

CISA-ADP ↗Reported

Currency of the record

CISA added this to its KEV catalog 15 days after the CVE record was published.KEV lag = CISA-KEV dateAdded − CVE.org datePublished; a floor on how long the known-exploited flaw sat in the public record before the federal catalog listed it. Severity/weakness here is CNA- or CISA-assigned (current at disclosure), independent of NVD's analysis backlog.

CISA KEV ↗CVE.org ↗Reported

WeaknessCWE-22 · Path TraversalPath traversal / file

Who’s exploiting it?

— attribution turns risk into urgency

Attribution not established

No confirmed (advisory-backed) threat-actor attribution is established for this record. Absence of a named actor is not absence of compromise — see Coverage & confidence.

Why it matters

— the attack path, told twice: adversary, then board

Attacker/business narrative framing is pending triage for this record — see Coverage & confidence below. Facts and verdict above are derived deterministically from the public feed.

What to do

— defensible action

Remediate per the vendor advisory — confirm the fixed build for your version and verify exposure.¹

Say it to the boardA vulnerability with this evidence profile is a defensible budget line, not a backlog ticket — fund the change against the proof above.

Coverage & confidence

— what we know, and what we don’t

Established (cited)

KEV listing (CISA)

EPSS probability (FIRST)

2 reported-exploitation source(s)

CVSS 5.3 — CNA-assigned (CVE.org)

CISA SSVC triage disposition (CVE.org ADP)

CWE weakness mapping — CNA-assigned (CVE.org)

Catalogued by JFROG (CNA)

Named finder/reporter credit (CVE.org)

Coverage gaps — stated, not hidden

No EUVD / GCVE mirror in feed — single-authority dependency for the identifier.

EPSS & exposure are time-varying; verify live at the source.

Threat-actor attribution not established from feed data — absence of a name is not absence of compromise.

Attacker/business narrative framing pending triage.

Disclosure & credit²

Catalogued by JFROGCNA

Credited with finding itKostya Kortchinsky | OpenAIfinder

Found an error? Propose a correction.

A correction is a cited counter-claim — it must cite a source, gets reviewed by a second human, and is never auto-applied. See the correction convention and our identity npub1j7gt9ky7sfcrjn8jjee6693dkzvk4rahs0fk2suu7968dfns62zs3mqm3y.

With JavaScript off: publish a signed kind 1111 Nostr event tagging ["r", "https://basicsecurity.net/threat/cve-2026-66384.html"], ["v", "sha256:723fdb2d21c9aadc5ca30eab49c94e192db3a3496c3f1582103511365cfc1463"] (the record version), plus ["src", "<your source URL>"] and ["t", "correction"], per the correction convention. Our identity: npub1j7gt9ky7sfcrjn8jjee6693dkzvk4rahs0fk2suu7968dfns62zs3mqm3y. Cite-or-die: an event with no src is a coverage finding, not a proposal.