Marco Cova

Saner: Composing Static and Dynamic Analysis to Validate Sanitization in Web Applications

Web applications are ubiquitous, perform mission- critical tasks, and handle sensitive user data. Unfortunately, web applications are often implemented by developers with limited security skills, and, as a result, they contain vulnerabilities. Most of these vulnerabilities stem from the lack of input validation. That is, web applications use malicious input as part of a sensitive operation, without having properly checked or sanitized the input values prior to their use. Past research on vulnerability analysis has mostly focused on identifying cases in which a web application directly uses external input in critical operations. However, little research has been performed to analyze the correctness of the sanitization process. Thus, whenever a web application applies some sanitization routine to potentially malicious input, the vulnerability analysis assumes that the result is innocuous. Unfortunately, this might not be the case, as the sanitization process itself could be incorrect or incomplete. In this paper, we present a novel approach to the analysis of the sanitization process. More precisely, we combine static and dynamic analysis techniques to identify faulty sanitization procedures that can be bypassed by an attacker. We implemented our approach in a tool, called Saner, and we applied it to a number of real-world applications. Our results demonstrate that we were able to identify several novel vulnerabilities that stem from erroneous sanitization procedures.

Swaddler: an approach for the anomaly-based detection of state violations in web applications

In recent years, web applications have become tremendously popular, and nowadays they are routinely used in security-critical environments, such as medical, financial, and military systems. As the use of web applications for critical services has increased, the number and sophistication of attacks against these applications have grown as well. Most approaches to the detection of web-based attacks analyze the interaction of a web application with its clients and back-end servers. Even though these approaches can effectively detect and block a number of attacks, there are attacks that cannot be detected only by looking at the external behavior of a web application. In this paper, we present Swaddler, a novel approach to the anomaly-based detection of attacks against web applications. Swaddler analyzes the internal state of a web application and learns the relationships between the applications critical execution points and the applications internal state. By doing this, Swaddler is able to identify attacks that attempt to bring an application in an inconsistent, anomalous state, such as violations of the intended workflow of a web application. We developed a prototype of our approach for the PHP language and we evaluated it with respect to several real-world applications.

Multi-module vulnerability analysis of web-based applications

In recent years, web applications have become tremendously popular, and nowadays they are routinely used in security-critical environments, such as medical, financial, and military systems. As the use of web applications for critical services has increased, the number and sophistication of attacks against these applications have grown as well. Current approaches to securing web applications focus either on detecting and blocking web-based attacks using application-level firewalls, or on using vulnerability analysis techniques to identify security problems before deployment. The vulnerability analysis of web applications is made difficult by a number of factors, such as the use of scripting languages, the structuring of the application logic into separate pages and code modules, and the interaction with back-end databases. So far, approaches to web application vulnerability analysis have focused on single application modules to identify insecure uses of information provided as input to the application. Unfortunately, these approaches are limited in scope, and, therefore, they cannot detect multi-step attacks that exploit the interaction among multiple modules of an application. We have developed a novel vulnerability analysis approach that characterizes both the extended state and the intended workflow of a web application. By doing this, our analysis approach is able to take into account inter-module relationships as well as the interaction of an applications modules with back-end databases. As a result, our vulnerability analysis technique is able to identify sophisticated multi-step attacks against the applications workflow that were not addressed by previous approaches. We implemented our technique in a prototype tool, called MiMoSA, and tested it on several applications, identifying both known and new vulnerabilities.

SNOOZE: toward a stateful network protocol fuzZEr

Fuzzing is a well-known black-box approach to the security testing of applications. Fuzzing has many advantages in terms of simplicity and effectiveness over more complex, expensive testing approaches. Unfortunately, current fuzzing tools suffer from a number of limitations, and, in particular, they provide little support for the fuzzing of stateful protocols. In this paper, we present SNOOZE, a tool for building flexible, security-oriented, network protocol fuzzers. SNOOZE implements a stateful fuzzing approach that can be used to effectively identify security flaws in network protocol implementations. SNOOZE allows a tester to describe the stateful operation of a protocol and the messages that need to be generated in each state. In addition, SNOOZE provides attack-specific fuzzing primitives that allow a tester to focus on specific vulnerability classes. We used an initial prototype of the SNOOZE tool to test programs that implement the SIP protocol, with promising results. SNOOZE supported the creation of sophisticated fuzzing scenarios that were able to expose real-world bugs in the programs analyzed.

ClearShot: Eavesdropping on Keyboard Input from Video

Eavesdropping on electronic communication is usually prevented by using cryptography-based mechanisms. However, these mechanisms do not prevent one from obtaining private information through side channels, such as the electromagnetic emissions of monitors or the sound produced by keyboards. While extracting the same information by watching somebody typing on a keyboard might seem to be an easy task, it becomes extremely challenging if it has to be automated. However, an automated tool is needed in the case of long-lasting surveillance procedures or long user activity, as a human being is able to reconstruct only a few characters per minute. This paper presents a novel approach to automatically recovering the text being typed on a keyboard, based solely on a video of the user typing. As part of the approach, we developed a number of novel techniques for motion tracking, sentence reconstruction, and error correction. The approach has been implemented in a tool, called ClearShot, which has been tested in a number of realistic settings where it was able to reconstruct a substantial part of the typed information.

Symbolic String Verification: An Automata-Based Approach

We present an automata-based approach for the verification of string operations in PHP programs based on symbolic string analysis. String analysis is a static analysis technique that determines the values that a string expression can take during program execution at a given program point. This information can be used to verify that string values are sanitized properly and to detect programming errors and security vulnerabilities. In our string analysis approach, we encode the set of string values that string variables can take as automata. We implement all string functions using a symbolic automata representation (MBDD representation from the MONA automata package) and leverage efficient manipulations on MBDDs, e.g., determinization and minimization. Particularly, we propose a novel algorithm for language-based replacement. Our replacement function takes three DFAs as arguments and outputs a DFA. Finally, we apply a widening operator defined on automata to approximate fixpoint computations. If this conservative approximation does not include any badpatterns (specified as regular expressions), we conclude that the program does not contain any errors or vulnerabilities. Our experimental results demonstrate that our approach works quite well in checking the correctness of sanitization operations in real-world PHP applications.

Analyzing and Detecting Malicious Flash Advertisements

The amount of dynamic content on the web has been steadily increasing. Scripting languages such as JavaScript and browser extensions such as Adobes Flash have been instrumental in creating web-based interfaces that are similar to those of traditional applications. Dynamic content has also become popular in advertising, where Flash is used to create rich, interactive ads that are displayed on hundreds of millions of computers per day. Unfortunately, the success of Flash-based advertisements and applications attracted the attention of malware authors, who started to leverage Flash to deliver attacks through advertising networks. This paper presents a novel approach whose goal is to automate the analysis of Flash content to identify malicious behavior. We designed and implemented a tool based on the approach, and we tested it on a large corpus of real-world Flash advertisements. The results show that our tool is able to reliably detect malicious Flash ads with limited false positives. We made our tool available publicly and it is routinely used by thousands of users.

Static Detection of Vulnerabilities in x86 Executables

Several approaches have been proposed to perform vulnerability analysis of applications written in high-level languages. However, little has been done to automatically identify security-relevant flaws in binary code. In this paper, we present a novel approach to the identification of vulnerabilities in x86 executables in ELF binary format. Our approach is based on static analysis and symbolic execution techniques. We implemented our approach in a proof-of-concept tool and used it to detect taint-style vulnerabilities in binary code. The results of our evaluation show that our approach is both practical and effective

Analysis of a Botnet Takeover

Botnets, networks of malware-infected machines (bots) that are controlled by an adversary, are the root cause of a large number of security problems on the Internet. A particularly sophisticated and insidious type of bot is Torpig, a malware program designed to harvest sensitive information (such as bank account and credit-card data) from its victims. In this article, the authors report on their efforts to take control of the Torpig botnet and study its operations for a period of 10 days. During this time, they observed more than 180,000 infections and recorded almost 70 Gbytes of data that the bots collected. They also report on what happened in the year that has passed since they lost control of the Torpig botnet.

An analysis of rogue AV campaigns

Rogue antivirus software has recently received extensive attention, justified by the diffusion and efficacy of its propagation. We present a longitudinal analysis of the rogue antivirus threat ecosystem, focusing on the structure and dynamics of this threat and its economics. To that end, we compiled and mined a large dataset of characteristics of rogue antivirus domains and of the servers that host them. The contributions of this paper are threefold. Firstly, we offer the first, to our knowledge, broad analysis of the infrastructure underpinning the distribution of rogue security software by tracking 6,500 malicious domains. Secondly, we show how to apply attack attribution methodologies to correlate campaigns likely to be associated to the same individuals or groups. By using these techniques, we identify 127 rogue security software campaigns comprising 4,549 domains. Finally, we contextualize our findings by comparing them to a different threat ecosystem, that of browser exploits. We underline the profound difference in the structure of the two threats, and we investigate the root causes of this difference by analyzing the economic balance of the rogue antivirus ecosystem. We track 372,096 victims over a period of 2 months and we take advantage of this information to retrieve monetization insights. While applied to a specific threat type, the methodology and the lessons learned from this work are of general applicability to develop a better understanding of the threat economies.