cyberleelawat/CVE-2025-55182

GitHubRepository PoCExploit

PoC files

5 files

File viewing is interactive and short-lived. Downloads are password-protected ZIP archives using password eip.

GitHub
Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The primary artifact is a Python script (CVE-2025-55182.py) that constructs and sends a malicious multipart/form-data POST request to a target server. The payload exploits a prototype pollution vulnerability in React Flight Protocol deserialization to achieve remote code execution by injecting a JavaScript payload that uses Node.js child_process.spawnSync to execute an attacker-supplied command and exfiltrates the output via a crafted redirect error.

Backdoor review

No backdoor observed in reviewed code

The supplied evidence consists of a Python exploit script, a README, and a Nuclei template, all targeting CVE-2025-55182. The Python script sends a crafted multipart POST request to execute a user-supplied command on a vulnerable React server and retrieves output via the X-Action-Redirect header. The Nuclei template performs the same attack with a hardcoded 'id' command. The README documents usage and defensive scanning. No concealed, deceptive, or operator-directed harmful behavior beyond the stated exploit was observed in the reviewed text.

ClassificationExploit
Model confidence98%
AuthenticationNot required
Languagespythonyamlmarkdown
Target softwarereact-server-dom-webpackreact-server-dom-turbopackreact-server-dom-parcelNext.js
Attack typesremote code executionprototype pollutiondeserialization of untrusted data
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The Python script CVE-2025-55182.py is explicitly designed to execute arbitrary commands on a remote target. It builds a payload that injects and runs code via spawnSync (lines 45-50), sends it to a target (lines 120-127), and retrieves the command output (lines 134-146). This constitutes an exploit, not a scanner or writeup.

CVE-2025-55182.py:2-5CVE-2025-55182.py:45-50CVE-2025-55182.py:120-127CVE-2025-55182.py:134-146

Requirements

  • Target must be running a vulnerable version of React Server Components (19.0.0, 19.1.0, 19.1.1, 19.2.0) with an exposed Server Function endpoint.CVE-2025-55182.py:2-5README.md:15-27
  • The target server must accept multipart/form-data POST requests to the Server Function endpoint.CVE-2025-55182.py:113-127

Observed behavior

  • The script constructs a JSON payload containing a malicious _prefix string that executes a shell command via Node.js child_process.spawnSync and throws a custom error to exfiltrate the output.CVE-2025-55182.py:45-50
  • The script sends the payload as a multipart/form-data POST request to the target URL with a Next-Action header.CVE-2025-55182.py:99-127
  • The script parses the HTTP response, specifically looking for an X-Action-Redirect header containing the command output, and prints it to the console.CVE-2025-55182.py:134-146
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Exploit Mechanism
Prototype pollution via $1:__proto__:then and Function constructor access via $1:constructor:constructorThe payload structure in CVE-2025-55182.py (lines 53-63) and cve-2025-55182.yaml (line 46) uses these techniques to achieve RCE on the target server.CVE-2025-55182.py:53-63cve-2025-55182.yaml:46
Command Execution
spawnSync('sh',['-c','<command>'])The Python script (line 46) and Nuclei template (line 46) execute arbitrary shell commands on the target server.CVE-2025-55182.py:46cve-2025-55182.yaml:46
Output Exfiltration
NEXT_REDIRECT header with command output in login?a= parameterThe exploit retrieves command output by throwing a crafted error that sets the X-Action-Redirect header (CVE-2025-55182.py lines 48-49, 134-144; cve-2025-55182.yaml line 46, 104-105).CVE-2025-55182.py:48-49CVE-2025-55182.py:134-144cve-2025-55182.yaml:46cve-2025-55182.yaml:104-105
Review boundaries

What the analysis did not establish

  • The evidence packet reports complete_artifact_coverage as false; two files (scanner.sh and another unspecified file) are omitted from the text content, so the full repository content is not available for analysis.
  • The analysis is based solely on the provided source code text; the code was not executed, and its effectiveness or safety is not verified.
  • Two files (scanner.sh and one other) were present in the repository but not included in the text evidence; their content was not reviewed.
  • Binary content was flagged as metadata-only and not analyzed, though no binary files were identified in the artifact.
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.

Linked vulnerabilities

1