Exploit catalog results

Showing 2 PoCs on this page

GitHub

cazzysoci/cve-2026-48908

Repository PoCStars: 0Created 2026-07-09
ExploitCVE-2026-489083 files

55.4 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that exploits CVE-2026-48908 in SP Page Builder for Joomla by uploading a crafted ZIP file containing a PHP webshell via the unauthenticated custom icon upload endpoint, then checks for successful shell access.

Backdoor review

No backdoor observed in reviewed code

The PoC script exploits CVE-2026-48908 by uploading a ZIP containing a PHP webshell. The webshell provides file upload functionality on the target, which is the expected exploit behavior. No concealed backdoor, credential theft, persistence, or unrelated payload was observed.

ClassificationExploit
Model confidence98%
AuthenticationNot required
LanguagesPython
Target softwareSP Page Builder for Joomla
Attack typesRemote Code ExecutionUnrestricted File Upload
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively uploads a PHP webshell to a target server and verifies its execution, which constitutes exploitation rather than mere detection or scanning.

poc.py:75-104

Requirements

  • Target URL running vulnerable SP Page Builder extension (versions 1.0.0-6.6.1)poc.py:110-114
  • List of target URLs provided as input filepoc.py:139-140

Observed behavior

  • Constructs a ZIP archive containing a PHP webshell disguised as an icon font uploadpoc.py:31-49
  • Sends POST request to the vulnerable upload endpoint without authenticationpoc.py:51-58
  • Checks if the uploaded PHP shell is accessible and executable by verifying HTTP 200 response containing the marker 'Shinday'poc.py:60-67
  • Writes successfully exploited shell URLs to result.txtpoc.py:128-131
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Webshell
Payload withheldThe PHP payload embedded in the ZIP provides a file upload form on the compromised server. This is the intended exploit outcome for the CVE.poc.py:32
Exploit Mechanism
Payload withheldThe script uploads a crafted ZIP containing a PHP file to the vulnerable endpoint, consistent with the described arbitrary file upload vulnerability.poc.py:51-58poc.py:110-115
Review boundaries

What the analysis did not establish

  • Only one text file (poc.py) was provided; two additional files in the repository were omitted from the evidence packet.
  • The evidence packet reports complete_artifact_coverage as false, indicating not all repository content was included.
  • Two additional files in the repository (total 3 files) were not provided as text and were not analyzed. Their content is unknown, but the PoC script itself shows no backdoor behavior.
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

cazzysoci/cve-breeze-2026

Repository PoCStars: 0Created 2026-04-26
ExploitUnlinked4 files

Python · 46.9 KiB

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

A Python script that checks if a WordPress site runs a vulnerable version of the Breeze plugin (CVE-2026-3844) and, if so, attempts to upload a PHP webshell via the plugin's Gravatar upload functionality to achieve remote code execution.

Backdoor review

No backdoor observed in reviewed code

The supplied evidence is a single Python script (breeze.py) that claims to be a PoC exploit for CVE-2026-3844 against the WordPress Breeze plugin. The script checks the target plugin version, then attempts to upload a PHP webshell via a gravatar-related endpoint. The embedded PHP webshell provides standard file manager capabilities (browse, upload, edit, delete, download). The script's behavior is consistent with a straightforward exploit PoC: it does not contain concealed backdoor functionality, credential exfiltration, persistence mechanisms, or unrelated payloads. No obfuscation or deceptive behavior was observed.

ClassificationExploit
Model confidence98%
AuthenticationNot required
Languagespython
Target softwareWordPress Breeze Plugin
Attack typesremote code executionwebshell upload
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively uploads a PHP webshell to a remote target and verifies its accessibility, which constitutes exploitation, not just detection or scanning.

breeze.py:155-222

Requirements

  • Target must have a vulnerable version of the Breeze plugin installed (<=2.1.19 or <=2.4.4).breeze.py:34-35
  • The target's Breeze plugin must have the Gravatar upload feature enabled and accessible.breeze.py:155-222

Observed behavior

  • Checks the target's Breeze plugin version by fetching readme.txt or breeze.php and parsing the version string.breeze.py:126-145
  • Creates a local PHP webshell file (shell.php) containing file management and command execution capabilities.breeze.py:38-118breeze.py:160-162
  • Attempts to upload the webshell to the target via POST requests to multiple endpoints using the 'breeze_fetch_gravatar' action.breeze.py:168-201
  • Tests if the uploaded webshell is accessible at several predicted URLs and reports success if a 200 response is received.breeze.py:207-222
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Embedded Webshell
Payload withheldThe script embeds a PHP webshell that is written to a local file and uploaded to the target. The webshell is a standard file manager with no hidden or unrelated functionality.breeze.py:38-118
Exploit Behavior
Payload withheldThe exploit_gravatar_upload function attempts to upload the webshell to several WordPress endpoints, then checks for successful deployment at predictable URLs. This is normal exploit behavior.breeze.py:155-222
Review boundaries

What the analysis did not establish

  • Only the main Python script (breeze.py) was provided; 3 other files in the repository were omitted from the evidence.
  • The evidence does not include any output or logs confirming successful execution of the exploit.
  • Three additional files in the repository were not included in the text evidence; their content is unknown.
  • The analysis is limited to static review of the provided text; no dynamic execution or network analysis 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.