Skip to main content
Mallory
CriticalPublic exploitVulnerability

CVE-2023-48022: Unauthenticated Remote Code Execution in Ray Jobs API (ShadowRay)

Published Nov 28, 2023 · Updated today
Trending this weekNot trendingOutside this week’s top 100
Mention activity · 30 days1 mention · peak 1 in a dayToday
CVSS9.8Critical · v3.1
EPSS83.9%100th percentile
CISA KEVNoNot listed
In the wildReported183 exploitation records in Mallory
Vendor
Anyscale
IdentifiersCVE-2023-48022CWE-306

CVE-2023-48022 allows remote attackers to execute arbitrary code through Ray's unauthenticated job submission API, including in Ray 2.6.3 and 2.8.0. A client with network access to an unprotected submission interface can submit a job whose entrypoint specifies an attacker-controlled shell command. Execution occurs in the Ray job or worker security context, and Ray scheduling capabilities can distribute malicious workloads across cluster nodes. The vendor disputes the vulnerability classification because Ray is intended to execute trusted code within a strictly controlled network environment. Ray 2.52.0 and later offer optional token authentication; upgrading alone does not establish that this protection is enabled.

Do you run this?See affected versions in your environment and whether adversaries are exploiting this vulnerability against assets like yours.
Check your exposure
Analyst brief

Impact, mitigation & remediation

What it means and what to do now: patch path, interim mitigations, and the assume-compromise checklist.

Impact

What an attacker gets, and what they’ve been doing with it.

Successful exploitation compromises the confidentiality, integrity, and availability of resources accessible to the Ray execution account. Attackers can access mounted datasets, model weights, source code, cloud credentials, service-account tokens, and other workload secrets, and consume CPU and GPU resources. Cluster scheduling can extend malicious execution to additional nodes, including nodes without direct internet exposure. Observed exploitation includes cryptomining, reverse shells, credential and data theft, cross-cluster propagation, and DDoS activity. System-level control depends on the execution account's privileges or subsequent privilege escalation.

Mitigation

If you can’t patch tonight, do this now.

Restrict Ray dashboard access, normally on TCP port 8265, and job-agent interfaces using firewall rules, private network segments, cloud security groups, and authenticated access controls. Permit connections only from trusted clients and cluster peers; protecting only the dashboard is insufficient if a job agent remains exposed. Run workloads with least privilege and limit accessible credentials and mounted sensitive data. Monitor job submissions, unexpected shell processes, outbound connections, persistence changes, and unexplained CPU or GPU use. Validate exposure using Ray Open Ports Checker or equivalent network checks.

Remediation

Patch, then assume compromise.

No conventional patch or fixed release is identified for this disputed issue. Deploy Ray only within its intended strictly controlled network environment and remove untrusted access to all job submission interfaces. Where appropriate, upgrade to Ray 2.52.0 or later and explicitly configure optional token authentication. Treat authentication as an additional control rather than a replacement for network isolation, and verify that both dashboard and job-agent interfaces are protected.

Public exploits

Exploits

3 exploits validated by Mallory's analysis.

View more in app →
ray_agent_job_rce.rbThis repository contains a single Metasploit module (modules/exploits/linux/http/ray_agent_job_rce.rb) that exploits a remote code execution (RCE) vulnerability in the Ray agent (CVE-2023-48022). The exploit targets the job submission HTTP API endpoints (/api/jobs/ and /api/job_agent/jobs/) on the default port 8265, which by default are unauthenticated. The module checks the Ray version via /api/version and only proceeds if the version is <= 2.6.3. It supports multiple Linux architectures and can deliver both direct command execution and staged payloads (such as Meterpreter reverse shells) using the CmdStager functionality. The exploit is weaponized, as it is part of the Metasploit framework and allows for flexible payload delivery. The main attack vector is network-based, requiring access to the Ray agent's HTTP API. The repository is well-structured, with clear separation of functionality for version checking, job submission, and payload delivery.Disclosed Nov 15, 2023 · Framework: metasploit · ruby · networkWeaponizedShadowRay-RCE-PoC-CVE-2023-48022This repository provides a weaponized exploit for CVE-2023-48022, a remote command execution vulnerability in the Ray framework (version 2.8.0). The exploit targets Ray's job submission API, which by default is unauthenticated and allows remote attackers to submit arbitrary jobs for execution. The repository contains two main exploit files: a Python script (exploit.py) for direct exploitation and a Metasploit module (ray_job_rce.rb) for integration with the Metasploit framework. The Python script allows users to specify a target Ray cluster and a command to execute, submitting the job via the Ray JobSubmissionClient. The Metasploit module automates exploitation, supporting both custom command execution and reverse shell payloads, and targets the '/api/jobs/' and '/api/job_agent/jobs/' endpoints. The README provides detailed usage instructions for both approaches. The exploit is operational and weaponized, enabling remote, unauthenticated code execution on vulnerable Ray clusters accessible over the network.jakabakos · Disclosed Mar 29, 2024 · Framework: metasploit · python · ruby · networkWeaponizedCVE-2023-48022The repository contains a standalone Python exploit client and a deliberately vulnerable Python HTTP mock illustrating CVE-2023-48022, described as missing authentication on Ray job submission. The seven files comprise two Python scripts, a Dockerfile, Docker Compose configuration, README, MIT license, and gitignore. No exploit framework is used. Repository URL, revision, and archive size were not supplied; blank strings and zero represent unavailable metadata. poc/poc.py submits the hardcoded command 'x; touch /tmp/<marker>; echo ' as the JSON entrypoint field to POST /api/jobs. mock/mock.py accepts JSON or form data without authentication and interpolates the entrypoint into "sh -c '{v}'", executed through subprocess.run with shell=True and a five-second timeout. This genuinely permits arbitrary shell command execution in the mock, although the provided client only creates a marker. Output is discarded. The mock is a simplified vulnerability model, not an actual Ray installation; successful exploitation of real Ray is not demonstrated. Verification uses GET /__lab_status, a laboratory-only endpoint listing filenames with the fixed cve-2023-48022 prefix. Custom markers lacking that prefix can therefore produce false negatives, while pre-existing marker files can produce false positives. The client does not validate submission success before checking for markers. Its local-address safeguard uses string prefixes rather than validated address ranges. The container uses Python 3.12 Alpine, an unprivileged user, dropped capabilities, no-new-privileges, loopback-only host publishing, and a /tmp tmpfs. Despite a source comment claiming no network access, the supplied configuration does not disable container networking. There are no active external callbacks, credential theft, persistence mechanisms, or destructive payloads.krevetkqqq · Disclosed Oct 30, 2026 · python · network · webOperational
Exposure surface

Affected products & vendors

Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.

VendorProductType
AnyscaleRayapplication
FAQ

Frequently asked questions

Is CVE-2023-48022 exploited in the wild?
Yes. As of Nov 19, 2026, Mallory has reports of exploitation in the wild.
Is there a public exploit for CVE-2023-48022?
Yes. Mallory tracks 3 public exploits or proofs of concept for CVE-2023-48022, the earliest disclosed Nov 15, 2023.
Is there a patch for CVE-2023-48022?
No conventional patch or fixed release is identified for this disputed issue. Deploy Ray only within its intended strictly controlled network environment and remove untrusted access to all job submission interfaces.
What products are affected by CVE-2023-48022?
Anyscale Ray.
How severe is CVE-2023-48022?
CVE-2023-48022 has a CVSS v3.1 score of 9.8 (Critical) and an EPSS score of 83.9% (100th percentile) as of Nov 19, 2026.
Activity feed

Recent activity

32 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.

View more in app →
cysecurity newsNews · Nov 17, 2026New Linux Malware Uses Tor, Ethereum Blockchain, and Falsified GPU Data to Run Cryptomining Operations Undetected - CySecurity News - Latest Information Security and Hacking IncidentsA vulnerability in Ray exploited during the historical ShadowRay 2.0 campaign to submit malicious jobs and distribute cryptominers across cluster nodes. The article does not explicitly attribute the current SANSAR campaign to this CVE.Read more ↗chocapikkNews · Oct 17, 2026Xorbits Xoscar: Unauthenticated RCE via Cloudpickle Deserialization in the Actor Pool - Chocapikk's Cybersecurity BlogA Ray vulnerability cited as a comparison to the Xoscar flaw: an internal RPC surface that deserializes attacker-controlled bytes and becomes network-exposed in distributed deployments. This reference provides no specific exploitation, patch, or mitigation status for the Ray vulnerability.Read more ↗vulncheck blogNews · Oct 15, 2026Xorbits Xoscar: Unauthenticated RCE via Cloudpickle Deserialization in the Actor Pool | Blog | VulnCheckA Ray vulnerability cited only as a comparison to the Xoscar issue: an internal RPC surface accepts attacker-controlled serialized data and becomes network-accessible in distributed deployments. The content does not provide specific exploitation, patch, or mitigation information for this CVE.Read more ↗nuclei templates pull requestsNews · Sep 12, 2026[CVE-2025-62593] Ray - Remote Code Execution by kiyx · Pull Request #17232 · projectdiscovery/nuclei-templates · GitHubA CVE identifier mentioned without any accompanying technical details or stated relationship to the primary Ray vulnerability.Read more ↗
+15 more in the timeline. Sign up to see them
Do you run this?See affected versions in your environment and whether adversaries are exploiting this vulnerability against assets like yours.
Check your exposure
Mallory platform

The version that knows your environment.

This page is what’s public. Mallory adds the parts that aren’t: which of your assets are affected, which adversaries are exploiting it right now, and which detections to deploy tonight.
Exposure mappingQuery your assets running an affected version, and investigate the blast radius.
Threat actor evidence4Every report linking this CVE to a named adversary.
Associated malware15Malware families linked to this CVE, with evidence and IOCs.
Detection signatures2YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Vendor-by-vendor mappingCross-references every affected SKU, including bundled OEM variants.
Social activity12Community discussion across Reddit, Mastodon, and other social sources.