EUVD-2026-12635
GHSA-j3mh-qmjj-xp83
Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessLocate exposed Ray Dashboard on port 8265
-
DeliverySend HTTP request with ../ traversal sequence
-
ExploitStatic handler resolves path outside asset root
-
ExecutionServer returns arbitrary file contents
-
ImpactHarvest credentials and config for further compromise
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V5 - File Handling Path canonicalization, traversal prevention
MITRE ATT&CK
2 techniques across 2 tactics: Discovery, Collection The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires network reachability to the Ray Dashboard HTTP service on its default port 8265 running a Ray version below 2.8.1; no authentication, user interaction, or special privileges are needed (AV:N/AC:L/PR:N/UI:N). The concrete prerequisite is that the Dashboard's static file handler is accessible to the attacker - exploitation succeeds against default configurations once the dashboard endpoint is exposed. The primary limiting factor is exposure: many deployments bind the dashboard to internal or loopback interfaces, so attackers without access to the cluster's management network cannot reach the vulnerable handler, and the impact is confined to read-only file disclosure (no write, RCE, or availability impact). |
| Impact | Signals are mixed and should be weighed together rather than in isolation. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N/VC:H) describes a network-reachable, low-complexity, unauthenticated read of high-value data, which scores 8.7 and looks alarming on paper. However, the EPSS probability is very low at 0.07% (20th percentile), and there is no CISA KEV listing, suggesting that despite a public PoC, mass exploitation is not currently observed. The realistic determinant of exposure is reachability: the Ray Dashboard is frequently bound to internal networks and is not authenticated by default, so risk concentrates wherever port 8265 is exposed to untrusted networks or the internet. Technical impact is limited to confidentiality (VI:N/VA:N), and the attack is plausibly automatable given a simple HTTP request, so internet-exposed dashboards should be treated as high priority while purely internal, network-segmented deployments are lower priority. No conflicting data was identified between the CVSS vector and the description - both consistently indicate remote, unauthenticated, read-only access. |
| Exploit ScenarioDerived | An attacker who can reach an exposed Ray Dashboard on port 8265 sends a single crafted HTTP request to the static-file endpoint containing '../' traversal sequences targeting a sensitive file such as /etc/passwd, application source, or cloud credential files. Because no authentication is required (PR:N) and complexity is low (AC:L), the server returns the file contents directly, and the attacker can iterate to harvest configuration secrets for further compromise. Publicly available exploit code (referenced via PacketStorm) lowers the skill barrier to weaponize this. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv5 – Jun 30, 2026
Local file disclosure in the Ray Dashboard (default TCP port 8265) of the Ray distributed-computing framework affects all releases prior to 2.8.1, where the static file handler fails to sanitize user-supplied paths. Remote attackers can submit directory-traversal sequences to read arbitrary files outside the intended static directory, exposing source code, configuration, and credentials. Publicly available exploit code exists (reported by VulnCheck), though EPSS estimates exploitation probability at only 0.07% (20th percentile) and the issue is not listed in CISA KEV.
Earlier versions of this analysis (4)
Local file disclosure in Ray (the distributed-computing/ML framework) versions prior to 2.8.1 lets remote unauthenticated attackers read arbitrary files on the host running the Ray Dashboard. The dashboard's static-file handler (default TCP port 8265) fails to sanitize user-supplied paths, so directory-traversal sequences (../) escape the intended static directory. Publicly available exploit code exists (referenced via Packet Storm and a VulnCheck advisory), though the EPSS probability is low (0.07%, 20th percentile) and the CVE is not on the CISA KEV list.
Local file disclosure in Ray Dashboard (default port 8265) affects all Ray versions prior to 2.8.1, where the static file handler fails to sanitize user-supplied paths. Remote unauthenticated attackers can submit '../' traversal sequences to read arbitrary files outside the intended static directory, exposing secrets, configuration, and credentials on the host. Publicly available exploit code exists (PacketStorm), though EPSS remains low at 0.07% and the flaw is not listed in CISA KEV.
Local file disclosure in Ray Dashboard (default port 8265) for Ray versions before 2.8.1 lets remote unauthenticated attackers read arbitrary files outside the static directory by submitting ../ traversal sequences to the static file handler. Publicly available exploit code exists (PacketStorm), though EPSS rates exploitation probability low at 0.07% and the flaw is not on the CISA KEV list. The CVSS 4.0 score of 8.7 reflects high confidentiality impact with no authentication barrier.
A path traversal vulnerability in Ray Dashboard (default port 8265) allows unauthenticated remote attackers to read arbitrary files from the server filesystem by exploiting improper path validation in the static file handling mechanism. The vulnerability affects Ray versions prior to 2.8.1 and has a publicly available proof-of-concept exploit, making it a high-priority security risk for organizations using Ray for distributed computing. With a CVSS score of 7.5 and network-accessible attack vector requiring no authentication, this vulnerability poses an immediate threat to data confidentiality.
RemediationAI
Vendor-released patch: upgrade Ray to 2.8.1 or later, which is the first release containing the fix (all versions below 2.8.1 are vulnerable). On Red Hat-based systems, apply the corresponding errata (RHSA-2026:5809, RHSA-2026:6761, RHSA-2026:6762, RHSA-2026:19712, or RHSA-2026:24977 as applicable to your channel). Where immediate patching is not possible, the most effective compensating control is to remove network exposure of the Dashboard: bind port 8265 to localhost or a trusted management interface and place it behind a reverse proxy that enforces authentication and normalizes/rejects path-traversal sequences, accepting the trade-off that operators lose direct browser access to the dashboard. Restrict ingress to port 8265 with host or network firewall rules to known administrative sources, and avoid exposing the Ray cluster or dashboard directly to the internet; note these network controls do not fix the underlying flaw and must be lifted only after upgrading. Refer to the VulnCheck advisory (https://www.vulncheck.com/advisories/ray-dashboard-path-traversal-leading-to-local-file-disclosure) and NVD entry (https://nvd.nist.gov/vuln/detail/CVE-2026-32981) for authoritative details.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in pypi
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Ray is an open-source framework for scaling Python and AI/ML workloads across clusters; its Dashboard is a web service (default port 8265) that serves bundled static assets alongside live cluster telemetry. The flaw is a classic CWE-22 Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal'): the static-file route concatenates attacker-controlled path segments without canonicalizing or constraining them to the asset root, so sequences such as '../' resolve to locations elsewhere on the host filesystem. The affected component is identified by CPE cpe:2.3:a:ray-project:ray:*:*:*:*:*:*:*:* for all versions below 2.8.1. Because the handler only reads and returns file contents, the impact is confidentiality-only disclosure rather than write or code execution.
Know when Ray gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
Network-reachable dashboard needs no auth or interaction (AV:N/AC:L/PR:N/UI:N); read-only file disclosure gives C:H with no integrity/availability impact and no scope change.
Primary rating from Vendor (VulnCheck).
CVSS VectorVendor: VulnCheck
Lifecycle Timeline
9DescriptionCVE.org
A path traversal vulnerability was identified in Ray Dashboard (default port 8265) in Ray versions prior to 2.8.1. Due to improper validation and sanitization of user-supplied paths in the static file handling mechanism, an attacker can use traversal sequences (e.g., ../) to access files outside the intended static directory, resulting in local file disclosure.
Executive ViewAI
Ray distributed-computing framework versions prior to 2.8.1 contain a path-traversal vulnerability in the Dashboard service that allows remote attackers to read arbitrary files, including source code, configuration files, and credentials. Organizations using Ray with accessible dashboards face credential exposure and source-code theft risk.
Apply vendor patch (Ray 2.8.1+). Interim: restrict network access to Ray Dashboard (port 8265) using firewall rules to trusted networks only; disable Dashboard service if not operationally required; monitor logs for directory-traversal patterns (../ sequences in requests).
Threat Intelligence Overview
NVD
CVE.org
EPSS
VulDB
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Path traversal exploits occur when applications use user-controlled input to construct file system paths without proper validation.
Related Vulnerabilities in Ray Dashboard Path Traversal and DoS
View all 2Ray is an AI compute engine. Rated critical severity (CVSS 9.4), this vulnerability is remotely exploitable, no authenti
A command injection existed in Ray's cpu_profile URL parameter allowing attackers to execute os commands on the system r
Anyscale Ray 2.6.3 and 2.8.0 allows /log_proxy SSRF. Rated critical severity (CVSS 9.1), this vulnerability is remotely
Remote code execution in Ray (the distributed compute/ML framework) before version 2.56.0 lets attackers run arbitrary c
LFI in Ray's /static/ directory allows attackers to read any file on the server without authentication. Rated high sever
LFI in Ray's log API endpoint allows attackers to read any file on the server without authentication. Rated high severit
Ray dashboard versions 2.53.0 and below lack proper authentication on DELETE endpoints, allowing unauthenticated attacke
Exploit-DB