Active Exploitation
#threatreport #LowCompleteness
Artifactory Under Attack: In-the-Wild Exploitation of CVE-2026-42016, CVE-2026-42018 & CVE-2026-82329 | 11-09-2026
Source: https://www.wiz.io/blog/artifactory-under-attack-in-the-wild-exploitation-of-cve-2026-42016-cve-2026-4201
Key details below ↓
🎯Victims: Organizations running jfrog artifactory
🔓CVEs: CVE-2026-42018 \[[Vulners](https://vulners.com/cve/CVE-2026-42018)]
- CVSS V3.1: *7.5*,
- Vulners: Exploitation: Unknown
CVE-2026-82329 \[[Vulners](https://vulners.com/cve/CVE-2026-82329)]
- CVSS V3.1: *9.8*,
- Vulners: Exploitation: True
Soft:
- jfrog artifactory (<7.111.21, <7.117.28, <7.125.20, <7.133.29, <7.146.38)
CVE-2026-42016 \[[Vulners](https://vulners.com/cve/CVE-2026-42016)]
- CVSS V3.1: *Unknown*,
- Vulners: Exploitation: Unknown
📚TTPs:
⚔️Tactics: 2
🛠️Technics: 0
🤖LLM extracted TTPs:`
T1059, T1059.004, T1068, T1078, T1083, T1105, T1136, T1190, T1505, T1505.003, ...
🧨IOCs:
- IP: 19
- Url: 2
- Hash: 1
- Domain: 0
- Email: 0
- BrowserExtension: 0
🔢Algorithms: sha1
📜Programming Languages: groovy, rust
💻Platforms: intel
#threatreport:
Between August 15 and September 8, 2026, threat actors actively exploited CVE-2026-42018 and CVE-2026-42016 against self-hosted JFrog Artifactory instances. CVE-2026-42018 allows an unauthenticated requester to obtain a JWT for Artifactory’s internal anonymous user, even when anonymous access is disabled. CVE-2026-42016 insufficiently validates token scope, allowing a low-privileged token to be exchanged or used with elevated administrative authority. Chained together, the vulnerabilities enable an unauthenticated attacker to obtain an admin-scoped token.
Observed attacks began with a trailing-slash request to `POST /access/api/v1/aws/token/`, which returned HTTP 200 and an anonymous-user JWT. The attacker then submitted the token to `POST /access/api/v1/tokens`, receiving an admin-scoped token that retained the username `anonymous`. Subsequent actions appeared in logs as originating from `token:anonymous`. In some cases, attackers created a persistent administrator account within five minutes using `PUT /api/security/users/<username>` or `PUT /access/api/ui/users/<username>`.
Post-exploitation activity included deploying malicious Groovy plugins through Artifactory’s native plugin framework to execute arbitrary code. Attackers used plugin endpoints such as `GET` or `POST /api/plugins/execute/<plugin>` to run shell commands, conduct reconnaissance, and enumerate files. They also delivered second-stage payloads by downloading binaries over HTTP, writing them to world-writable locations including `/dev/shm`, `/tmp`, and `/var/tmp`, and establishing command-and-control communication. A custom Rust-based backdoor was observed in multiple incidents, while some actors uploaded scripts to repository paths to maintain webshell-style access.
From September 1 to September 8, 2026, separate actors exploited the critical CVE-2026-82329, an authentication-bypass vulnerability affecting Artifactory’s default configuration. Exploitation involved unauthenticated requests to `POST /access/api/v1/registry/join`, which returned HTTP 201 and an admin-scoped token. Subsequent activities varied across environments. At initial disclosure, approximately 67–69% of observed Artifactory deployments were vulnerable to the three flaws. Remediation remained incomplete, although exposure to CVE-2026-82329 declined more rapidly because of its critical severity.
Post summary
The report documents real‑world exploitation of three Artifactory CVEs, detailing how attackers obtain admin‑level tokens, deploy malicious Groovy plugins, and establish persistence, underscoring an active threat landscape.