Skip to main content
Mallory
CriticalPublic exploit

Unauthenticated RCE in Anyscale Ray Job Submission API

IdentifiersCVE-2023-48022CWE-306

CVE-2023-48022 affects Anyscale Ray 2.6.3 and 2.8.0 and stems from missing authentication on the Ray dashboard/job submission interface. When the Ray dashboard API is exposed, a remote attacker can submit jobs through the unauthenticated Jobs API (commonly referenced as /api/jobs/ on the dashboard, default dashboard port 8265) and cause arbitrary code to execute on the Ray cluster. Reporting and incident writeups describe attackers abusing the exposed dashboard API to run multi-stage Bash and Python payloads, leveraging Ray orchestration features to execute across cluster nodes and pivot to additional internal systems. The vendor position noted in the source material is that Ray is intended only for strictly controlled network environments; later versions reportedly offer optional token authentication, but the issue is described in the provided content as historically unpatched for the affected versions.

Share:
For your environment

Are you exposed to this one?

Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.

ANALYST BRIEF

Impact, mitigation & remediation

What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.

Impact

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

Successful exploitation enables full remote code execution on exposed Ray clusters. In observed intrusions, attackers used this access to hijack CPU/GPU resources for XMRig cryptomining, deploy reverse shells, establish persistence, move laterally to non-internet-facing nodes, steal sensitive assets including AI models, datasets, source code, cloud and database credentials, and retained production data, and repurpose compromised infrastructure for DDoS activity. On large GPU-backed environments, the operational and financial impact can be substantial due to compute theft and broader cluster compromise.

Mitigation

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

Primary mitigation is strict network isolation: deploy Ray only inside trusted, controlled networks and block public access to the dashboard and job submission endpoints. Place the service behind authenticated access controls, bastions, or private networking, and add authorization on top of the Ray dashboard where possible. Continuously monitor for anomalous job submissions, unexpected Bash/Python execution, reverse shells, miner processes, unauthorized cron/systemd changes, and lateral execution across nodes. If public exposure cannot be eliminated immediately, tightly limit source IPs and disable or restrict job submission functionality until compensating controls are in place.

Remediation

Patch, then assume compromise.

Do not expose the Ray dashboard or Jobs API to the public internet. Upgrade to a Ray release that supports token authentication and enable that control where available; the provided content states customers on version 2.52.0 and later can choose to use token authentication. Restrict access to the Ray dashboard/API using firewall rules, security groups, VPN-only access, or equivalent network segmentation so only trusted administrative hosts can reach it. Review cluster configurations for accidental exposure, including the default dashboard port 8265, and use vendor tooling such as the Ray Open Ports Checker referenced in the content to validate exposure. Because exploitation has been observed in the wild, incident response should include credential rotation, review of scheduled tasks/cron/systemd persistence, and inspection of all cluster nodes for malicious jobs and payloads.
PUBLIC EXPLOITS

Exploits

1 valid exploit after Mallory filtered fakes, detection scripts, and README-only repos.

VALID 1 / 1 TOTALView more in app
ShadowRay-RCE-PoC-CVE-2023-48022MaturityPoCFrameworkmetasploitVerified exploit

This 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.

jakabakosDisclosed Mar 29, 2024pythonrubynetwork
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

Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.

ACTIVITY FEED

Recent activity

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

What this page doesn’t show

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, which detections to deploy, and what to do tonight.
Exposure mapping

Query your assets running an affected version, and investigate the blast radius.

Threat actor evidence2

Every observed campaign linking this CVE to a named adversary.

Associated malware7

Malware families riding this exploit, with evidence and IOCs.

Detection signatures1

YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.

Vendor-by-vendor mapping

Cross-references every affected SKU, including bundled OEM variants.

Social activity12

Community discussion across Reddit, Mastodon, and other social sources.