Update README.md

Fiex

README.md

@@ -1,2 +1,65 @@
-# keylogger-detector
-University project for an Operating Systems lecture. The goal is to develope a keystroke-logger-detector for a Linux environment. Developement Environment: Fedora 37 VM under  Gnome on VirtualBox. A project journal can be found [here](https://github.com/sebaschi/keylogger-detector/blob/main/doc/dev_journal.md)
+# KLDetect
+KLDetect is a keylogger detector for the Linux Desktop.
+It can detect processes reading from ```/dev/input/event*``` devices and kernel modules registered to listen to keyboard events.
+
+# Dependencies
+* [Python](https://www.python.org/downloads/)
+* [SystemTap](https://sourceware.org/systemtap/wiki)
+* [```fuser```](https://www.man7.org/linux/man-pages/man1/fuser.1.html)
+* Utilities that come with [Fedora](https://fedoraproject.org/) like ```which```.
+
+# Setup
+Download or clone this repository:
+```
+git clone https://github.com/sebaschi/keylogger-detector.git
+```
+Navigate into the src directory:
+```
+cd keylogger-detector/src
+```
+Run a keylogger. KLDetect has been tested and shown to work on the following keylogger.
+
+User progams:
+* [simple-key-logger](https://github.com/gsingh93/simple-key-logger/tree/master)
+* [logkeys](https://github.com/kernc/logkeys)
+* [keylog](https://github.com/SCOTPAUL/keylog)
+
+
+Kernel Module:
+* [spy](https://github.com/jarun/spy)
+
+# Usage 
+KLDetect **must** be run as root (sudo). 
+If KLDetect is not run as root the user is reminded to try again with root permissions.
+
+Running without options just runs userspace detection:
+```
+./kldetect.py
+```
+To get a list of command line options:
+```
+./kldetect.py -h
+```
+To run with kernel module detection:
+```
+./kldetect.py -k
+```
+To run just kernel module detection
+```
+./kernel_detector.py
+```
+
+# Warning
+Running any part if this program in a lightheaded manner may break your system.
+Killing processes and unloading modules should be done with caution. We suggest testing it an a VM.
+If one runs the KLDetect with the kernel module keylogger detection option set, make sure to update the [whitelist.txt](https://github.com/sebaschi/keylogger-detector/blob/main/src/whitelist.txt) with the safe kernel modules that you know you have on your system. In particular we highly suggest running ```lsmod > <path-to-kldetect>/whitelist.txt```, before inserting a kernel keylogger. This writes the modules currently inserted in the kernel to the whitelist. This way 'normal' modules that you already have installed on the 'clean' kernel will not accidentally be unloaded. Altough KLDetect should not unload any kernel modules currently used, better safe than sorry.
+
+# Developers
+Copyright © 2023[Michel Romancuk](https://github.com/SoulKindred), [Sebastian Lenzlinger](https://github.com/sebaschi)
+
+
+
+
+
+This project is Part of a Univeristy project at the [Operating Systems](https://dmi.unibas.ch/de/studium/computer-science-informatik/lehrangebot-fs23/vorlesung-operating-systems-1/) lecture at the University of Basel, Switzerland.
+ A project journal can be found [here](https://github.com/sebaschi/keylogger-detector/blob/main/doc/dev_journal.md).

doc/dev_journal.md

@@ -166,7 +166,7 @@ This is after extensivly refactoring because I was starting to loose oversight o
 ## Wednesday, 7. June 2023, day
 ### Sebastian
 VirtualBox stopped working so after much pain I decided to switch to Boxes. There the install of Fedora 37 went smoothly.
-Then Started testing the userland detector on [simple-key-logger](https://github.com/gsingh93/simple-key-logger/tree/maste), [logkeys](https://github.com/kernc/logkeys).
+Then Started testing the userland detector on [simple-key-logger](https://github.com/gsingh93/simple-key-logger/tree/master), [logkeys](https://github.com/kernc/logkeys).
 [pykeylogger](https://github.com/amoffat/pykeylogger) produced a segmentation fault, after I finaly got it to run. Trying to run [py-keylogger](https://github.com/hiamandeep/py-keylogger), turns out it only runs on X11 it seem (so we'd not catch it anyway).
 [keylog](https://github.com/SCOTPAUL/keylog) was succesfully detected and removed.
 All in all, the main functionality works as intended. Basically now would be the refinement phase to add more options or to have a way to configure the config.json file more easily.

doc/research/acm_2390317.2390326.bib

@@ -0,0 +1,17 @@
+@inproceedings{10.1145/2390317.2390326,
+author = {Howard, Adam and Hu, Yi},
+title = {An Approach for Detecting Malicious Keyloggers},
+year = {2012},
+isbn = {9781450315388},
+publisher = {Association for Computing Machinery},
+address = {New York, NY, USA},
+url = {https://doi.org/10.1145/2390317.2390326},
+doi = {10.1145/2390317.2390326},
+abstract = {Keyloggers are applications that are installed onto computers with the intent of monitoring and storing keystrokes that are input by a user. These keystrokes can either be stored on a physical hard disk or transmitted via a network connection to a remote location. Because of their functions, keyloggers have a potential of being used for malicious purposes. In order to protect privacy, it is important to realize the threat that a keylogger application might pose and identify appropriate methods for detecting it. The method presented in this research provides a standardized approach to detect unknown keylogging software from a computer. We also conducted experiments on a variety of keyloggers to verify the effectiveness of the proposed approach.},
+booktitle = {Proceedings of the 2012 Information Security Curriculum Development Conference},
+pages = {53–56},
+numpages = {4},
+keywords = {rootkit, privacy, system hook, keylogger, malicious software},
+location = {Kennesaw, Georgia},
+series = {InfoSecCD '12}
+}

doc/research/acm_financial_losses_due_to_malware.bib

@@ -0,0 +1,17 @@
+@inproceedings{10.1145/2905055.2905362,
+author = {Amin, Maitri},
+title = {A Survey of Financial Losses Due to Malware},
+year = {2016},
+isbn = {9781450339629},
+publisher = {Association for Computing Machinery},
+address = {New York, NY, USA},
+url = {https://doi.org/10.1145/2905055.2905362},
+doi = {10.1145/2905055.2905362},
+abstract = {General survey stat that the main damage malware can cause is to slow down their PCs and perhaps crash some websites which is quite wrong, The Russian antivirus software developer teamed up with B2B International for a study worldwide recently, shown 36\% of users lose money online as a result of a malware attack. Currently malware can't be detected by traditional way based anti-malware tools due to their polymorphic and/or metamorphic nature. Here we have improvised a current detection technique of malware based on mining Application Programming Interface (API) calls and developed the first public dataset to promote malware research.• In survey of cyber-attacks 6.2\% financial attacks are due to malware which increase to 1.3 \% in 2013 compared to 2012.• Financial data theft causes 27.6\% to reach 28,400,000. Victims abused by this targeting malware countered 3,800,000, which is 18.6\% greater than previous year.• Finance-committed malware, associated with Bitcoin has demonstrated the most dynamic development. Where's, Zeus is still top listed for playing important roles to steal banking credentials.Solutionary study stats that companies are spending a staggering amount of money in the aftermath of damaging attack: DDoS attacks recover $6,500 per hour from malware and more than $3,000 each time for up to 30 days to moderate and improve from malware attacks. [1]},
+booktitle = {Proceedings of the Second International Conference on Information and Communication Technology for Competitive Strategies},
+articleno = {145},
+numpages = {4},
+keywords = {Malware, API, financial losses, Survey},
+location = {Udaipur, India},
+series = {ICTCS '16}
+}

doc/research/acm_risk_of_stolen_credentials.bib

@@ -0,0 +1,17 @@
+@inproceedings{10.1145/3133956.3134067,
+author = {Thomas, Kurt and Li, Frank and Zand, Ali and Barrett, Jacob and Ranieri, Juri and Invernizzi, Luca and Markov, Yarik and Comanescu, Oxana and Eranti, Vijay and Moscicki, Angelika and Margolis, Daniel and Paxson, Vern and Bursztein, Elie},
+title = {Data Breaches, Phishing, or Malware? Understanding the Risks of Stolen Credentials},
+year = {2017},
+isbn = {9781450349468},
+publisher = {Association for Computing Machinery},
+address = {New York, NY, USA},
+url = {https://doi.org/10.1145/3133956.3134067},
+doi = {10.1145/3133956.3134067},
+abstract = {In this paper, we present the first longitudinal measurement study of the underground ecosystem fueling credential theft and assess the risk it poses to millions of users. Over the course of March, 2016--March, 2017, we identify 788,000 potential victims of off-the-shelf keyloggers; 12.4 million potential victims of phishing kits; and 1.9 billion usernames and passwords exposed via data breaches and traded on blackmarket forums. Using this dataset, we explore to what degree the stolen passwords---which originate from thousands of online services---enable an attacker to obtain a victim's valid email credentials---and thus complete control of their online identity due to transitive trust. Drawing upon Google as a case study, we find 7--25\% of exposed passwords match a victim's Google account. For these accounts, we show how hardening authentication mechanisms to include additional risk signals such as a user's historical geolocations and device profiles helps to mitigate the risk of hijacking. Beyond these risk metrics, we delve into the global reach of the miscreants involved in credential theft and the blackhat tools they rely on. We observe a remarkable lack of external pressure on bad actors, with phishing kit playbooks and keylogger capabilities remaining largely unchanged since the mid-2000s.},
+booktitle = {Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security},
+pages = {1421–1434},
+numpages = {14},
+keywords = {keylogger, phishing, risk analysis, data breach, password reuse, authentication, phishing kit, password},
+location = {Dallas, Texas, USA},
+series = {CCS '17}
+}

doc/research/citation-strange-world-keyloggers.bib

@@ -0,0 +1,10 @@
+@article{article,
+author = {Creutzburg, Reiner},
+year = {2017},
+month = {01},
+pages = {139-148},
+title = {The strange world of keyloggers - an overview, Part I},
+volume = {2017},
+journal = {Electronic Imaging},
+doi = {10.2352/ISSN.2470-1173.2017.6.MOBMU-313}
+}

src/config.json

@@ -1,8 +1,8 @@
 {
     "white_listed_programs": [
         "systemd_logind",
-        "gnome-shell",
-        "systemd"
+        "systemd",
+        "gnome-shell"
     ],
     "auto_kill_programs": [],
     "kbd_names": [

src/funcall_trace1.stp

@@ -1,8 +0,0 @@
-
-probe kernel.function("register_keyboard_notifier").call
-	{
-		printf("%s (%d)\n", execname(),tid())
-	}
-probe end{
-	printf("end\n")
-}

src/funcall_trace2.stp

@@ -1,58 +0,0 @@
-#! /usr/bin/env stap
-#
-# Copyright (C) 2010-2015 Red Hat, Inc.
-# Written by William Cohen <wcohen@redhat.com>
-#
-# The linetimes.stp script takes two arguments: where to find the function
-# and the function name. linetimes.stp will instrument each line in the
-# function. It will print out the number of times that the function is
-# called, a table with the average and maximum time each line takes,
-# and control flow information when the script exits.
-#
-# For example all the lines of the do_unlinkat function:
-#
-# stap linetimes.stp kernel do_unlinkat
-
-global calls, times, last_pp, region, cfg
-
-probe $1.function(@2).call { calls <<< 1 }
-probe $1.function(@2).return {
-  t = gettimeofday_us()
-  s = times[tid()]
-  if (s) {
-    e = t - s
-    region[last_pp[tid()]] <<< e
-    cfg[last_pp[tid()], pp()] <<< 1
-  }
-  delete times[tid()]
-  delete last_pp[tid()]
-}
-
-probe $1.statement(@2 "@*:*") {
-  t = gettimeofday_us()
-  s = times[tid()]
-  if (s) {
-    e = t - s
-    region[last_pp[tid()]] <<< e
-    cfg[last_pp[tid()], pp()] <<< 1
-  }
-  times[tid()] = t
-  last_pp[tid()] = pp()
-}
-
-probe end {
-  printf("\n%s %s call count: %d\n", @1, @2, @count(calls));
-  printf("\n%-58s %10s %10s\n", "region", "avg(us)", "max(us)");
-  foreach (p+ in region) {
-    printf("%-58s %10d %10d\n", p, @avg(region[p]), @max(region[p]));
-  }
-
-  printf("\n\ncontrol flow graph information\n")
-  printf("from\n\tto\n=======================\n")
-  foreach ([src+] in region) {
-     printf("%-s\n", src)
-     foreach ([s,dest+] in cfg[src,*]) { # slice for all dest's
-        printf("\t%-s %d\n", dest, @count(cfg[src,dest]));
-     }
-  }
-}

src/kernel_detector.py

@@ -1,3 +1,5 @@
+#!/usr/bin/env python3
+
 import subprocess
 import time
 import multiprocessing
@@ -136,43 +138,42 @@ def detect_logger(module):
 #
 #==============================================================================================================
 
-def run_kernel_detection:
-	whitelist = get_whitelist("whitelist.txt")
+whitelist = get_whitelist("whitelist.txt")
 
-	lsmod_output = list_modules("lsmod");
+lsmod_output = list_modules("lsmod");
 
-	sus_modules = compare_mods(whitelist, lsmod_output)
+sus_modules = compare_mods(whitelist, lsmod_output)
 
-	sus_modules = tidy_up(sus_modules)
+sus_modules = tidy_up(sus_modules)
 
-	sus_modules = unload_mod(sus_modules)
-	time.sleep(1)
+sus_modules = unload_mod(sus_modules)
+time.sleep(1)
 
-	sus_modules = getpath(sus_modules)
-	print(sus_modules)
-	if len(sus_modules) == 0:
-		print("nothing to do")
-		print("ALL CLEAN")
-		exit()
-	
-	
-	suspects = []
-	for module in sus_modules:
-		suspects.append(detect_logger(module))
-		time.sleep(1)
+sus_modules = getpath(sus_modules)
+print(sus_modules)
+if len(sus_modules) == 0:
+    print("nothing to do")
+    print("ALL CLEAN")
+    
+
+
+suspects = []
+for module in sus_modules:
+    suspects.append(detect_logger(module))
+    time.sleep(1)
 
 
 
-	print("Following modules are logging your keystrokes: ")
-	for i in range(len(suspects)):
-		print( f"[{i}] {suspects[i]}")
-	print("Enter the number of the module you want to remove: ")
-	user_input = input().split()
-	for j in user_input:
-		to_remove = suspects[int(j)]
-		subprocess.Popen(['sudo','rmmod', to_remove])
-		print(f"Removed {to_remove}")
-	print("Finished")
+print("Following modules are logging your keystrokes: ")
+for i in range(len(suspects)):
+    print( f"[{i}] {suspects[i]}")
+print("Enter the number of the module you want to remove: ")
+user_input = input().split()
+for j in user_input:
+    to_remove = suspects[int(j)]
+    subprocess.Popen(['sudo','rmmod', to_remove])
+    print(f"Removed {to_remove}")
+print("Finished")
 	
 	
 

src/kldetect.py

@@ -1,5 +1,7 @@
 #!/usr/bin/env python3
-
+import subprocess
+import time
+import os
 import sys
 from config import CONFIG_FILE, load_config, save_config
 from utils import (
@@ -22,6 +24,8 @@ safe_option = False
 add_white_list_option = False
 add_black_list_option = False
 debug_option = False
+kernel_detection_option = False
+
 
 # Functions
 def debug(option, to_print):
@@ -35,9 +39,10 @@ def print_help():
     print('  -v, --verbose\t\t\tVerbose mode. Informative information will be displayed duting execution')
     print('  -a, --auto-kill\t\tAutomatically kill blacklisted processes')
     print('  -s, --safe\t\t\tSafe mode. Asked to confirm before killing a process')
-    print('  -w, --add-white-list\t\t\tActivate prompt to add program names to the whitelist') #For some reason this line gets messed up in display 
-    print('  -b, --add-black-list\t\t\tAutomatically add program names chosen to kill to the blacklist')
+    print('  -w, --add-white-list\t\tActivate prompt to add program names to the whitelist') #For some reason this line gets messed up in display 
+    print('  -b, --add-black-list\t\tAutomatically add program names chosen to kill to the blacklist')
     print('  -d, --debug\t\t\tDebug mode. Print debug statements')
+    print(' -k, --kernel-detection\t\tRun the kernel keylogger detector, too. CURRENTLY NOT IMPLEMENTED TO DIRECTLY RUN KERNEL DETECTOR.')
 
 def set_input_options():
     """
@@ -50,7 +55,7 @@ def set_input_options():
     """
 
     global auto_kill_option, verbose_option, safe_option, add_white_list_option
-    global debug_option, add_black_list_option
+    global debug_option, add_black_list_option, kernel_detection_option
     if len(sys.argv) > 1:
         for arg in sys.argv[1:]:
             print(arg)
@@ -69,6 +74,8 @@ def set_input_options():
                 add_black_list_option = True
             elif arg == '-d' or arg == '--debug':
                 debug_option = True
+            elif arg == '-k' or arg == '--kernel-detection':
+                kernel_detection_option = True
    
 
 def confirm_kill_procces(process_name, times=0):
@@ -101,6 +108,176 @@ def confirm_kill_procces(process_name, times=0):
     else:
         return confirm_kill_procces(process_name, times+1)
 
+
+def detect_kernel(module):
+        """
+        Start the systemtap-script.
+        load and unload modules twice.
+        load module when finished.
+        
+        Args:
+            module(str): Path + name of the module being tested
+        
+        Returns:
+            String: Path + name of the module that is logging keystrokes  
+        """        
+        if verbose_option:
+            print('[Verbose] Started kernel keylogger detection')
+        process = subprocess.Popen(['stap','funcall_trace.stp', '-T', '10'], stdout=subprocess.PIPE, text=True)
+
+
+        for i in range(2):
+            subprocess.Popen(['sudo','insmod', module])
+            time.sleep(1)
+            print(".", end="")
+            subprocess.Popen(['sudo','rmmod', module])
+            time.sleep(1)
+        subprocess.Popen(['sudo','insmod', module])
+        print(".")
+        out = process.communicate()[0]
+        if verbose_option:
+            print('[Verbose] Started kernel keylogger detection')
+
+        print(out)
+        if out == "[-]":
+       	    return module
+        print("FAILED")
+        return 0
+		
+def getpath(sus_modules):
+	"""
+	Gets the path of a list of modules being tested
+	calls "find_file()" function
+	
+	Args:
+	    List[module(str)] List of all modules being tested
+	    
+	Returns:
+	    List[modules(str)]List of the Path of all modules being tested
+	"""
+	for i in range(len(sus_modules)):
+		sus_modules[i] = find_file(sus_modules[i] + ".ko")
+	return sus_modules
+		
+def find_file(filename):
+    """
+    Searches for a file begining at root
+    
+    Args:
+        filename(str) The filename one is looking for
+        
+    Returns:
+        result_out(str) 'The Path_to_Module/Module_name'
+    """
+    result = []
+    for root, dirs, files in os.walk("/"):
+        if filename in files:
+            file_path = os.path.join(root, filename)
+            result.append(file_path)
+    result_out = result
+    result_out = ''.join(result_out)
+    return result_out
+
+def unload_mod(modules):
+	"""
+	Unloads modules.
+	
+	Args:
+	    module(str) the module that needs to be unloaded. Has to be Path_to_Module/Module_name
+	"""
+	tmp = []
+	for module in modules:
+		result = subprocess.run(['sudo','rmmod', module],capture_output = True, text = True)
+		if result.returncode == 0:
+			if verbose_option:
+				print(f"[Verbose] Unloaded module: {module}")
+		else:
+			if verbose_option:
+				print(f"[Verbose] Failed to unloaded module: {module}")
+				print("[Verbose] " + result.stderr)
+			tmp.append(module)
+	result_out = compare_mods(tmp, modules)
+	if verbose_option:
+		print("[Verbose] ", end="")
+		print(result_out)
+	return result_out
+		
+		
+def tidy_up(entries):
+	"""
+	Takes a txt file and removes everything except the first word of a line
+	
+	Args:
+	    File(.txt) in this usecase a whitelist.txt
+	    
+	Returns:
+	    clean_entries(List[str]) List of only the first wrod from each line
+	"""
+	cleaned_entries = []
+	for entry in entries:
+		modules = entry.split()
+		if modules:
+			first_mod = modules[0]
+			cleaned_entries.append(first_mod)
+	return cleaned_entries
+		
+def compare_mods(A, B):
+	"""
+	Does set-suptraction to.
+	
+	Args:
+	    A(list[str]) List of elements one wants to ignore 
+	    B(list[str]) List of elements that one wants without all elements in A
+	    
+	Returns:
+	    result(list[str] List of elements that are in B but not in A
+	"""
+	setA = set(A)
+	setB = set(B)
+	
+	result = setB - setA
+	
+	return list(result)
+
+
+def get_whitelist(file_path):
+	"""
+	reads a text-file
+	
+	Args:
+	    file_path(str) Path to file one wants to read
+	    
+	Returns:
+	    lines(list[str]) List of each line from a file
+	"""
+	try:
+		with open(file_path, 'r') as file:
+			lines = file.read().splitlines()
+			return lines
+	except IOError:
+		print(f'Error: Failed to load whitelist{file_path}')
+			
+def list_modules(command):
+	"""
+	Calls a command in terminal
+	
+	Args:
+	    command(str) the command one wants to execute
+	    
+	Returns:
+	    result(list[std]) List of each line the command has as an output.
+	"""
+
+	result = subprocess.run(command, shell = True, capture_output=True, text=True)
+	
+	if result.returncode == 0:
+		return result.stdout.strip().split('\n')
+	else:
+		print(f"Failed with error:{result.stderr}")
+		return[]
+
+
+
 def detect_keyloggers():
     """
     Detect (userland) keylogger processes based on which processes have a keyboard file open (/dev/input/event*)
@@ -116,7 +293,7 @@ def detect_keyloggers():
     debug(True, str(sys.argv)) # Set manually to debug if args are being read
     check_platform()
 
-    global auto_kill_option, verbose_option, safe_option
+    global auto_kill_option, verbose_option, safe_option, add_white_list_option, kernel_detection_option, debug_option
     set_input_options()
     if verbose_option:
         print('[Verbose] Input options set')
@@ -205,7 +382,8 @@ def detect_keyloggers():
     ############################
     if len(suspicious_processes) == 0:
         print("[+] No suspicious processes found")
-        sys.exit(0)
+        if not kernel_detection_option:
+            exit(0)
 
     ############################
     # 7. Prompt user to chose which processes (not covered by auto kill if set) to kill
@@ -282,11 +460,60 @@ def detect_keyloggers():
     if verbose_option:
         print('[Verbose] Config file saved')
     
-    print('[+] Program completed. Exiting.')
+    
+
+
+   
+    debug(debug_option, 'Kernel detection option: ' + str(kernel_detection_option))
+    
+    ###########################
+    # 10. If kernel_detection_option is set, run kernel detection
+    ###########################
+    
+    
+    if kernel_detection_option:
+        whitelist = get_whitelist("whitelist.txt")
+        lsmod_output = list_modules("lsmod")
+        sus_modules = compare_mods(whitelist, lsmod_output)
+        sus_modules = tidy_up(sus_modules)
+        sus_modules = unload_mod(sus_modules)
+        time.sleep(1)
+        sus_modules = getpath(sus_modules)
+        suspects = []
+        if verbose_option:
+            print("[Verbose] ", end="")
+            print(sus_modules)
+        if len(sus_modules) == 0 and verbose_option:
+            print("[Verbose] Nothing to do")
+		    
+        for module in sus_modules:
+            if module == '': #if modules have an empty path, they are in root
+        	    break
+            suspects.append(detect_kernel(module))
+            time.sleep(1)
+
+        print("Following modules are logging your keystrokes: ")
+        for i in range(len(suspects)):
+            print( f"[{i}] {suspects[i]}")
+        print("Enter the number of the module you want to remove: ")
+        user_input = input().split()
+        to_remove = []
+        for j in user_input:
+            to_remove = suspects[int(j)]
+            subprocess.Popen(['sudo','rmmod', to_remove])
+        if len(to_remove) < 1:
+            print(f"Removed {to_remove}")
+        
+    print('[+] Program completed. Exiting.')		
+		
+    
+    
+
+
 
 if __name__ == '__main__':
     detect_keyloggers()
-
+