Exploit catalog results

Showing 6 PoCs on this page

GitHub

Cosm3No1de/SmartHire---Hack-The-Box-WriteUp

Repository PoCStars: 0Created 2026-07-23
WriteupCVE-2024-370543 files

HTML · 85.9 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

A detailed penetration testing walkthrough for the Hack The Box 'SmartHire' machine, documenting the exploitation of CVE-2024-37054 (MLflow pickle deserialization) and a Python .pth file hijacking privilege escalation. The artifact contains no executable exploit or scanner code; it is a technical report with redacted commands and payloads.

Backdoor review

No backdoor observed in reviewed code

The supplied evidence consists of a Hack The Box writeup (README.md and index.html) for the 'SmartHire' machine. The content describes a standard penetration testing methodology, including reconnaissance, exploitation of CVE-2024-37054 (MLflow pickle deserialization), and privilege escalation via Python .pth file hijacking. All commands and payloads are consistent with the documented attack path and target the stated victim machine. No concealed executable behavior, deceptive payloads, or operator-directed harm unrelated to the described CTF scenario was observed.

ClassificationWriteup
Model confidence100%
AuthenticationRequired
LanguagesEnglishSpanish
Target softwareMLflowPythonNginxOpenSSH
Attack typesDeserialization of Untrusted DataPrivilege Escalation
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact is a writeup. It provides a substantive technical analysis of a penetration testing engagement, including reconnaissance, exploitation, and privilege escalation steps. It does not contain functional exploit or scanner code; the provided code snippets are illustrative and redacted. The README.md explicitly states it is a 'complete penetration testing walkthrough' and the index.html is a styled presentation of the same content.

README.md:3-5README.md:136-138

Requirements

  • Access to the target MLflow instance with default credentials (admin:password).README.md:101-106
  • Ability to upload a malicious pickle payload to the MLflow model registry.README.md:111-121
  • Write access to the /opt/tools/mlflow_ctl/plugins/dev/ directory for privilege escalation.README.md:198-201

Observed behavior

  • Describes using default credentials to authenticate to an MLflow instance.README.md:101-106
  • Describes creating a malicious pickle payload with a __reduce__ method to execute a reverse shell command.README.md:126-131
  • Describes triggering the deserialization by sending a POST request to the /predict endpoint.README.md:140-143
  • Describes using a bind shell to bypass an egress firewall.README.md:159-162
  • Describes privilege escalation by creating a malicious .pth file and a fake mlflow_actions.py module to execute commands as root via sudo.README.md:218-235
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Exploit Payload
Payload withheldA reverse shell payload used to gain initial access as part of the documented CVE-2024-37054 exploitation.README.md:130
Exploit Payload
Payload withheldA bind shell payload used to bypass an egress firewall, consistent with the described privilege escalation path.README.md:162
Privilege Escalation Command
Payload withheldA command used in the malicious Python module to set the SUID bit on /bin/bash for root privilege escalation.README.md:226
Review boundaries

What the analysis did not establish

  • The artifact contains redacted IP addresses, credentials, and flags, which limits the ability to verify the exact commands used.
  • The 'exploit.py' script mentioned in the writeup is not included in the provided evidence.
  • The evidence includes a non-text media file (img3.png) that was not analyzed.
  • One non-text media file (img3.png) was present but not analyzed; it is a banner image and unlikely to contain executable code.
  • The analysis is limited to the supplied text files; no dynamic analysis or execution was performed.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

GitHub

Cosm3No1de/Bedside.htb_solved

Repository PoCStars: 0Created 2026-07-22
WriteupCVE-2025-645123 files

81.0 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a detailed educational writeup describing the step-by-step methodology used to solve the Hack The Box 'Bedside' machine. It documents the attack chain, including initial foothold via payload withheld (pdfminer.six pickle deserialization), user access via a Vite path traversal, and root escalation via torch.load() deserialization. It contains no executable exploit or scanner code; it only describes the techniques and provides illustrative code snippets.

Backdoor review

No backdoor observed in reviewed code

The artifact is a CTF writeup repository containing a README.md and an index.html file. Both files describe the exploitation of payload withheld and other techniques against a Hack The Box machine. The content is educational documentation with no executable code, no concealed payloads, and no instructions for the reviewer to perform harmful actions. The repository does not contain any backdoor or deceptive behavior.

ClassificationWriteup
Model confidence98%
AuthenticationNot required
LanguagesSpanishHTMLCSS
Target softwarepdfminer.sixVitePyTorchMONAI
Attack typesInsecure DeserializationPath TraversalPrivilege Escalation
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact is a writeup. It provides a substantive technical analysis of the attack chain used to compromise a machine, including vulnerability descriptions and code snippets, but it does not contain a functional, self-contained exploit or scanner. The code snippets are illustrative examples within the narrative, not standalone tools.

README.md:1-116index.html:1-1369

Requirements

  • Access to the Hack The Box 'Bedside' machine network.README.md:50-52
  • Ability to upload a malicious PDF and pickle file to the target web application.index.html:1076-1078

Observed behavior

  • Describes creating a malicious pickle file that executes a reverse shell command.index.html:1085-1092
  • Describes using a path traversal in a PDF to load the malicious pickle file via pdfminer.six.index.html:1096-1101
  • Describes using a Vite path traversal to read a developer's SSH private key.index.html:1122-1126
  • Describes creating a malicious PyTorch checkpoint to create a SUID bash binary for root escalation.index.html:1161-1168
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Cve Reference
Payload withheldThe writeup documents exploitation of this CVE for educational purposes.README.md:26index.html:1072-1073
Ip Address
Payload withheldExample reverse shell IP address in educational pickle payload code snippet.index.html:1088
Ip Address
Payload withheldExample SSH connection IP address in educational code snippet.index.html:1136
Review boundaries

What the analysis did not establish

  • The evidence consists of a README.md and an index.html file, which are the rendered writeup. No standalone exploit or scanner scripts were provided.
  • The analysis is based solely on the text content; no code was executed to verify the described techniques.
  • One non-text media file (img2.png) was present but not analyzed, as per the evidence policy.
  • One non-text file (img2.png, 28189 bytes) was present in the repository but not analyzed; its content is unknown.
  • The review is limited to the supplied text evidence; no dynamic analysis or execution was performed.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

GitHub

Cosm3No1de/HTB-Snapped-Writeup

Repository PoCStars: 0Created 2026-07-15
WriteupCVE-2026-27944CVE-2026-38883 files

HTML · 64.5 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

A Hack The Box writeup for the 'Snapped' machine, describing the exploitation chain of payload withheld (Nginx UI unauthenticated backup disclosure) and payload withheld (snapd TOCTOU local privilege escalation). The artifact contains no exploit or scanner code; it is a narrative technical analysis with command examples.

Backdoor review

No backdoor observed in reviewed code

The artifact is a Hack The Box writeup describing exploitation of payload withheld and payload withheld. The two text files (README.md and index.html) contain only educational documentation, static HTML/CSS, and shell command examples. No executable payload, obfuscated code, or concealed operator-directed harm is present in the reviewed text.

ClassificationWriteup
Model confidence100%
AuthenticationNot required
LanguagesSpanishEnglish
Target softwareNginx UIsnapd
Attack typesinformation disclosureprivilege escalation
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact is a writeup. It provides a substantive technical analysis of the exploitation chain for two CVEs, including reconnaissance, exploitation steps, and privilege escalation. It contains no exploit or scanner code; the commands shown are illustrative examples within a narrative, not functional code intended to be executed.

README.md:1-195index.html:1-374

Requirements

  • Access to the target network service on port 80README.md:43-47
  • Local user access on the target system for privilege escalationREADME.md:105

Observed behavior

  • Describes downloading an unauthenticated backup from Nginx UI and extracting the AES key from the X-Backup-Security headerREADME.md:61-69
  • Describes decrypting the backup, extracting a bcrypt hash, and cracking it to obtain SSH credentialsREADME.md:72-94
  • Describes compiling and executing a SUID exploit for payload withheld to achieve root privileges via a race condition in snapdREADME.md:107-156
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Cve Reference
Payload withheldThe writeup documents exploitation of this Nginx UI vulnerability for foothold.README.md:28index.html:241-243
Cve Reference
Payload withheldThe writeup documents exploitation of this snapd privilege escalation vulnerability.README.md:29index.html:244-247
External Repository Reference
Payload withheldThe writeup instructs cloning this repository to compile the exploit, but the artifact itself does not contain or execute that code.README.md:110index.html:350
Review boundaries

What the analysis did not establish

  • The artifact is a writeup describing an exploitation chain; it does not contain functional exploit or scanner code.
  • The evidence includes only the README.md and index.html files; a non-text media file (snapped.png) was not analyzed.
  • The writeup references external exploit repositories (e.g., TheCyberGeek/payload withheld) but does not include their source code.
  • One non-text file (snapped.png, 48292 bytes) was present in the artifact but was not analyzed; its content is limited to metadata in this review.
  • The review covers only the two text files supplied; any other files in the repository snapshot are not included in the evidence packet.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

GitHub

Cosm3No1de/TwoMillion-HTB-Write-up

Repository PoCStars: 0Created 2026-07-08
WriteupCVE-2023-03863 files

96.8 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

A Hack The Box writeup for the 'TwoMillion' machine. It describes a multi-step attack chain including API abuse, command injection, and privilege escalation via payload withheld. It contains no executable exploit or scanner code; it is a narrative technical analysis with example commands.

Backdoor review

No backdoor observed in reviewed code

The artifact is a Hack The Box write-up for the TwoMillion machine. It contains educational text, HTML, and CSS describing a penetration testing methodology, including API enumeration, command injection, and privilege escalation via payload withheld. No backdoor, trojan, or deceptive payload targeting the reader was observed. The included C code snippet is a non-functional, commented outline of an exploit, and the instructions reference cloning a well-known public repository. No concealed executable behavior or operator-directed harm is present.

ClassificationWriteup
Model confidence100%
AuthenticationRequired
LanguagesMarkdownHTMLCSSBashC
Target softwareLinux KernelWeb Application (API)
Attack typesCommand InjectionPrivilege EscalationInformation Disclosure
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact is a Hack The Box writeup. It provides a step-by-step narrative of an attack, including example curl commands and shell interactions, but does not contain functional exploit or scanner code. The included C code snippet is a non-functional placeholder. The primary purpose is technical analysis and documentation, not automated exploitation or scanning.

README.md:1-5README.md:306-329

Requirements

  • Access to the target web application API endpointsREADME.md:49-55
  • Ability to register a user account on the targetREADME.md:84-92
  • Local access to a vulnerable Linux kernel (5.15.70) for payload withheldREADME.md:188-194

Observed behavior

  • Describes using a PUT request to /api/v1/admin/settings/update to escalate a user to admin by setting is_admin=1README.md:96-102
  • Describes command injection via the username parameter in a POST to /api/v1/admin/vpn/generate, resulting in remote code execution as www-dataREADME.md:115-135
  • Describes lateral movement to the admin user by reading database credentials from a .env file and using SSHREADME.md:160-176
  • Describes privilege escalation to root by compiling and executing a public payload withheld exploitREADME.md:196-210
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Cve Reference
Payload withheldThe write-up describes using a public kernel exploit for privilege escalation on a target machine. This is standard educational content for a CTF write-up.README.md:24README.md:184-194
External Repository Reference
Payload withheldThe write-up instructs the reader to clone a public GitHub repository containing a known payload withheld proof-of-concept. This is a reference to an external resource, not a payload within the artifact.README.md:199index.html:598
Non Executable Code Snippet
Payload withheldThe artifact includes a commented, non-functional C code skeleton that describes the logic of an OverlayFS exploit. It is not a complete, compilable, or executable payload and serves only as an illustration.README.md:306-329
Review boundaries

What the analysis did not establish

  • The artifact contains a non-functional, summarized C code snippet (lines 309-329) that does not constitute a working exploit.
  • The evidence includes only the writeup text; no actual exploit binaries or scanner scripts are present.
  • The analysis is based solely on the provided text files; the non-text media file (twomillion.png) was not inspected.
  • One non-text media file (twomillion.png, 64067 bytes) was present but not analyzed per binary policy. It is used as a banner image in the HTML and is unlikely to contain executable content.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

GitHub

Cosm3No1de/HTB-Abducted-Writeup

Repository PoCStars: 0Created 2026-07-06
WriteupCVE-2026-44803 files

HTML · 61.2 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a HackTheBox machine write-up for 'Abducted'. It documents the steps to exploit CVE-2026-4480 (Samba RCE) and escalate privileges, but does not contain executable exploit or scanner code.

Backdoor review

No backdoor observed in reviewed code

The artifact is a HackTheBox write-up for the 'Abducted' machine. It documents exploitation of CVE-2026-4480 and privilege escalation steps. The reviewed text files (README.md and index.html) contain only educational documentation, commands, and references. No concealed executable behavior, deceptive payloads, or operator-directed harm is present in the supplied evidence.

ClassificationWriteup
Model confidence100%
AuthenticationNot required
LanguagesMarkdownHTML
Target softwareSamba
Attack typesOS command injectionprivilege escalation
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact is a write-up documenting the exploitation process for a HackTheBox machine. It contains no executable exploit or scanner code; it only describes commands and steps. The README.md and index.html files are purely descriptive and educational.

README.md:1-128index.html:1-358

Requirements

  • Network access to SMB shares with null authenticationREADME.md:24-25
  • Writable SMB printer share (HP-Reception)README.md:25

Observed behavior

  • Describes using a public PoC for CVE-2026-4480 to obtain a reverse shell as user 'nobody'README.md:27-29
  • Describes decoding rclone-obfuscated credentials to gain SSH access as user 'scott'README.md:33-35
  • Describes abusing SMB wide links and force user to escalate to user 'marcus'README.md:42-45
  • Describes escalating to root by writing a systemd drop-in to set SUID on /bin/bashREADME.md:48-53
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

URL
Payload withheldExternal PoC repository referenced in the write-up; not part of the reviewed artifact.README.md:77index.html:268
URL
Payload withheldAuthor's portfolio link in the HTML footer.index.html:351
Review boundaries

What the analysis did not establish

  • One non-text file (abducted.png) was not analyzed; it is a banner image and unlikely to contain code.
  • The artifact references an external PoC (payload withheld) but does not include its source code.
  • One non-text media file (abducted.png, 44636 bytes) was flagged as metadata-only and not analyzed; its content could theoretically contain steganographic payloads, but no evidence suggests this.
  • The artifact references an external PoC repository (TheCyberGeek/CVE-2026-4480-PoC) which is not included in the reviewed evidence and was not analyzed.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

GitHub

Cosm3No1de/htb-orion-writeup

Repository PoCStars: 0Created 2026-06-27
WriteupCVE-2025-32432CVE-2026-240616 files

780.1 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

A Hack The Box machine writeup describing the attack chain for the 'Orion' box, which involves exploiting CVE-2025-32432 (CraftCMS RCE) and CVE-2026-24061 (telnetd auth bypass). The artifact contains no exploit or scanner code; it is a narrative walkthrough with example commands.

Backdoor review

No backdoor observed in reviewed code

The artifact is a Hack The Box writeup repository containing a README.md and an index.html. Both files describe a penetration testing walkthrough for the 'Orion' machine, detailing exploitation of CVE-2025-32432 and CVE-2026-24061. The content is purely educational documentation with no executable code, concealed payloads, or instructions that would cause harm to the reader. No backdoor, trojan, or suspicious behavior is present.

ClassificationWriteup
Model confidence100%
AuthenticationNot required
LanguagesMarkdownHTMLPython
Target softwareCraftCMSGNU Inetutils telnetd
Attack typesRemote Code ExecutionAuthentication BypassCredential Harvesting
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact is a Hack The Box writeup containing a narrative description of an attack chain, example commands, and screenshots. It does not contain any executable exploit or scanner code; it is a technical analysis document.

README.md:1-296index.html:1-688

Requirements

  • Target running CraftCMS 5.6.16README.md:47
  • Target running GNU inetutils telnetd 2.7README.md:212

Observed behavior

  • Describes using Metasploit module exploit/linux/http/craftcms_preauth_rce_cve_2025_32432 to gain a reverse shellREADME.md:128-135
  • Describes using USER="-f root" telnet -a 127.0.0.1 to bypass authentication and gain root accessREADME.md:226-228
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Educational Content
Payload withheldThe repository is a static writeup with no executable components; it documents known vulnerabilities for educational purposes.README.md:1-296index.html:1-688
Review boundaries

What the analysis did not establish

  • Evidence consists of a README.md and an index.html file, both of which are writeup documents. No exploit or scanner code is present.
  • The artifact references CVE-2025-32432 and CVE-2026-24061 but does not contain original exploit code for either vulnerability.
  • Four non-text media files (767,599 bytes) were flagged as metadata-only and not analyzed; they are not part of the selected text evidence.
  • The review covers only the two selected text files; other repository files were omitted from the evidence packet.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.