With regulatory frameworks like GDPR in Europe and CCPA in California, a cybersecurity professional needs not only technical skills but also an understanding of legal constraints. This multidisciplinary approach expands their toolset, making them invaluable assets in the compliance process and, ultimately, in maintaining a company's reputation.
A common fallacy many fall into, including decision-makers in organizations, is the belief that cybersecurity measures are a one-time investment. This often arises from a need for more understanding of the velocity at which the technological landscape evolves. For the public, the lifetime of a tech stack may seem long, but for engineers, anything beyond a certain timeframe, say 12-15 months, often becomes obsolete or vulnerable. Using technologies like TLS 1.1, SHA-1, or even older versions of HTTPS can expose systems to substantial risks despite having been considered secure not too long ago.
There's a gap between organizational timelines and the relentlessness of cybercriminals. The latter aren't confined by budget quarters or product cycles. Many stakeholders are shocked when they realize that a security audit from three months ago doesn't make their systems impervious to new attacks. Security measures depreciate not in years but in months, sometimes weeks, depending on the threat vector. The advent of zero-day exploits and their quick proliferation amongst malicious actors compounds this issue. Saying, "But we just did X three months ago," is akin to believing that an expired vaccine can still provide immunity against an evolving virus.
Python is a linchpin for many of these varied skills thanks to its ease of use and versatility. Cybersecurity professionals can write quick scripts to automate routine tasks, develop complex machine learning models for predictive threat analysis, or even create customized tools for unique security requirements.
The beauty of Python lies in its ability to bridge the gap between various cybersecurity disciplines. For example, a security expert focusing on data encryption can use the same language and tools as a network analyst working on real-time traffic monitoring. This unification enables more cohesive team dynamics and a streamlined approach to cybersecurity issues.
Python’s role in cybersecurity
Python's ubiquity in the cybersecurity landscape is no accident; it results from a potent combination of features that make it the go-to language for security experts. The language boasts an expansive ecosystem of libraries and tools specifically tailored for cybersecurity tasks. From PyCrypto for cryptographic algorithms to Scapy for packet sniffing and manipulation, Python eliminates the need for reinventing the wheel, allowing professionals to focus on crafting advanced, custom solutions. The sheer volume of available packages often means that if you have a specific need, there's likely a library for it; if not, you have the community support to develop one.
And it's not just the libraries; it's the people behind them. With an enthusiastic community of developers contributing to forums like Stack Overflow and an increasing number of machine learning models like ChatGPT offering specialized support, the wealth of communal knowledge is astounding. This community-driven dynamic results in continuously updated official and unofficial documentation, ensuring that even the most complex tasks have some level of guidance available online.
Python is also the epitome of ease of use. Its straightforward syntax and readability mean that experts can quickly write scripts for automated tasks, saving precious time in high-stakes situations like incident response. For those new to cybersecurity, Python is a gentle introduction to programming concepts, making it easier to grasp complex topics like memory management, data structures, or algorithms. This ease of use also positions Python as a stepping stone for learning more specialized languages. Once you're comfortable with Python, transitioning to languages like C, Rust, or Go for specific performance or system-level tasks becomes significantly easier.
Automated threat analysis
Automated threat analysis has evolved beyond simple signature-based detection methods. In today's context, machine learning classifiers like Support vector machines (SVMs), Decision trees, Random forests, Neural networks, and Ensemble methods are at the forefront of identifying and predicting potential cyber threats.
- Support vector machines (SVMs) classify data points by finding the hyperplane that best divides the data classes. In a cybersecurity context, SVMs are often employed in anomaly detection where classifying network behavior as 'normal' or 'anomalous' is crucial.
- Decision trees break down a complex decision-making process into a more straightforward structure. In cybersecurity, they can be used to create rule-based classifiers for intrusion detection or to build decision-making pathways for automated responses to specific threats.
- Random forests are an extension of Decision Trees and work by creating a 'forest' of trees and making them 'vote' for a class. This reduces the overfitting problem seen in individual decision trees. Given their robustness, they are commonly used in classifying malware types based on their features.
- Neural networks, particularly deep learning architectures, excel in picking up intricate patterns from data. This is exceptionally beneficial in identifying previously unknown threats based on behavioral traits rather than known signatures.
- Ensemble methods like AdaBoost or Gradient Boosting are meta-algorithms that combine the predictions from multiple machine learning algorithms to improve performance. These are particularly effective when false positives are costly, as they can harmonize results from various classifiers to improve the overall prediction reliability.
Python’s most significant advantage is its ability to handle large datasets efficiently. This is indispensable when dealing with enormous network data, logs, or event-related information commonly associated with threat analysis.
- Scalability and flexibility: Libraries like Pandas allow Python to scale from small datasets to big data efficiently. Whether operating on a single machine or a cluster, Python provides the flexibility to handle data of varying sizes without significant changes to the codebase.
- Data manipulation and preprocessing: Python libraries like NumPy and Scikit-learn offer robust data manipulation capabilities. Preprocessing steps like normalization, encoding, or imputation can be done effortlessly, ensuring the dataset is optimally prepared for machine learning algorithms.
- Parallel computing: For ultra-large datasets, Python's compatibility with similar computing technologies like Hadoop and Spark allows it to distribute data processing tasks across multiple machines, further enhancing its ability to manage large datasets.
- Integrative ecosystem: The machine learning libraries (e.g., TensorFlow, PyTorch, Scikit-learn) seamlessly integrate with data manipulation libraries, making Python a one-stop-shop for data ingestion, manipulation, modeling, and analysis.
Malware analysis
In the labyrinthine cybersecurity landscape, malware analysis is akin to forensic science – meticulous, detail-oriented, and crucial for understanding the enemy. Python, with its versatile toolkit, has emerged as an indispensable asset for cybersecurity professionals engaged in malware research.
Python can be wielded as a scalpel to dissect malware down to its most granular elements. Libraries like PEfile allow for intricate analysis of Portable Executable (PE) files, commonly used in Windows executables. PEfile can parse the intricacies of PE headers and sections, thereby aiding in unraveling obfuscation techniques or extracting embedded payloads. Similarly, YARA is invaluable in threat hunting and pattern recognition within binaries. This allows for creating bespoke rule sets to unearth hidden or obfuscated code segments that might elude standard detection mechanisms. For those keen on taking a deeper plunge, Angr can be employed for binary analysis to virtually execute and path-explore a malware sample, mapping out its execution flow and identifying areas that warrant scrutiny.
The beauty of Python in malware analysis is its aptitude for automation. When dealing with polymorphic or metamorphic malware that changes its code to evade detection, the speed and efficiency of automated analysis are invaluable. Python allows for stitching together different analytical tools in a cohesive workflow. Imagine automating the metadata extraction with PEfile, passing it through a YARA rule-set for pattern identification, and then using Angr for detailed function-level analysis – all scripted in Python to run in sequence or parallel. This level of automation accelerates the analysis process and minimizes the chance of human error, mainly when repetitive tasks are involved.
But what sets Python apart is its unparalleled adaptability, which lends itself to developing custom tools and scripts tailored to unique malware analysis requirements. Whether writing a script to automate sandbox submissions and collect results or crafting a specialized parser to interpret a malware's custom communication protocol, Python provides the flexibility to create highly specialized analysis instruments. For instance, you could use Socket programming features to mimic C2 servers for a specific malware family, enabling real-time monitoring of its communication patterns.
Moreover, Python's compatibility with machine learning libraries facilitates the development of predictive models that can be integrated into custom malware classification tools. So, not only can you identify known malware variants, but you also have the computational power to predict new, previously unseen malware types based on learned behaviors and attributes.
Don't overlook Python's standard library, which contains modules like os and sys for file and system operations, subprocess for interacting with the system shell, and hashlib for checksum and hash computations. These can be woven into any custom tool to enhance its capabilities further, allowing you to interact deeply with system-level operations and data manipulations.
Network security
Network security is not just about setting up firewalls or intrusion detection systems; it's a proactive, real-time monitoring and response domain that ensures the heartbeats of an organization – the data flow – are regular, unobstructed, and secure. Python plays a vital role in enhancing these real-time monitoring capabilities by seamlessly integrating with existing security platforms or by facilitating the development of bespoke monitoring and automation tools.
Python's asynchronous programming capabilities, leveraged through libraries like asyncio, make it an excellent choice for real-time network monitoring. It can efficiently handle multiple network sockets concurrently, making it feasible to monitor large-scale network traffic without lag. Whether packet sniffing using Scapy or conducting port scans with Nmap's Python bindings (python-nmap), Python facilitates an extensive range of networking tasks to be executed efficiently.
Moreover, the versatility of Python is highlighted when it integrates with existing security solutions. Several leading network security platforms offer Python SDKs or APIs to augment their capabilities. Cisco's SecureX, Microsoft Sentinel, and Palo Alto Networks' Cortex XSOAR are perfect examples. SecureX, for instance, offers orchestration capabilities that can be automated using Python to perform complex tasks like multi-source data correlation or automated threat response actions. Microsoft Sentinel's Python SDK can be utilized to automate data ingestion, alerting, and even response actions, making it an invaluable asset in an organization's SIEM strategy. Similarly, Cortex XSOAR can be integrated with Python scripts to automate threat intel collection, incident response, or even remediation actions.
Given the abundance of such integrative solutions, the skillset demand shifts unmistakably toward Python proficiency. Having a team member adept in Python – or a dedicated Python developer – is no longer a luxury; it's an urgency. They can leverage the existing security infrastructure to its maximum potential and develop custom Python scripts that fill the gaps in your security posture. Whether automating routine tasks like log monitoring and analysis using Python's powerful libraries like Pandas or more complex actions like building alerting mechanisms using its smtplib for email notifications or Twilio for SMS alerts, Python's broad standard library can cover these needs extensively.
Network security is as dynamic as the threats it aims to neutralize. Python equips security professionals with the agility, adaptability, and power of automation, setting them a class apart in their defensive capabilities. It's not just a programming language; in the hands of a skilled cybersecurity professional, Python becomes a robust shield, a proactive sensor, and a rapid responder – all rolled into one.
DevSecOps
DevSecOps represents the evolution of DevOps, ingraining security into the development and operations process rather than tacking it on as a separate phase. This is a critical shift as "security as a bolt-on" is increasingly antiquated. With DevSecOps, security becomes an integral part of the product lifecycle. With its flexibility, robust libraries, and ease of use, Python is a critical tool in achieving this amalgamation. Here's a breakdown of Python's roles in different aspects of DevSecOps.
In a DevSecOps environment, the need for speed is counterbalanced by the need for security. Python's Ansible library allows server provisioning and configuration management tasks to be automated. Coupling this with Python's capacity to manage cloud services via libraries like Boto for AWS gives you a powerful, centralized, and automated mechanism that can handle complex, multi-cloud architectures. This significantly reduces the margin of human error and ensures that each provisioned instance adheres to a hardened configuration baseline. Additionally, Python scripts can interact with Infrastructure as Code (IaC) tools like Terraform to dynamically create and manage resources that meet specific security benchmarks.
Automation in CI/CD pipelines is nothing new. However, adding security features within the pipeline takes it to the next level. Python interfaces seamlessly with Jenkins, GitLab, and other CI/CD tools to trigger builds, run tests, and even revert code if it doesn't meet specific security criteria. The Python-based tool Bandit can perform static code analysis during the build process, flagging vulnerabilities in real-time and providing developers with immediate feedback. This ensures that insecure code doesn't progress further down the pipeline, facilitating early remediation.
Automated security testing, powered by Python, can happen in parallel with other development activities. Tools like OWASP ZAP provide Python APIs, enabling you to integrate penetration tests directly into your DevSecOps pipeline. Using Python, you can also automate complex, multi-stage attacks to gauge how well your security controls withstand advanced threats. For code-level vulnerabilities, Python's Pyflakes or Pylint can scan your repository for issues like command injection or insecure deserialization. This offers the advantage of identifying vulnerabilities before the application even makes it to a staging environment, reducing remediation costs and time.
Python excels in data analytics, making it an excellent choice for parsing and analyzing security logs, network flows, and endpoint threat detection data. The Pandas library can handle large data sets and perform transformations to identify patterns or anomalies. Scikit-learn can provide machine learning algorithms to classify potential security incidents. For network data, Python's Scapy library can dissect packets, identify suspicious activity, and even interact with existing intrusion detection systems to add or modify rules dynamically. This enables businesses to create a security feedback loop where threat intelligence is continuously refined and integrated into the DevSecOps process.
Python as a language of cybersecurity tooling
In the ever-evolving panorama of cybersecurity, the tools you deploy are as good as the intelligence and adaptability they offer. With its inherent versatility, Python not only becomes a language of choice for scripting and automation but also serves as the backbone for specialized cybersecurity tools designed to protect, detect, and remediate. Let's delve into three such Python-centric tools that testify to Python's indomitable stature in cybersecurity.
Bandit is a Python-born tool specializing in static code analysis, serving as a sentinel that continuously scans your Python codebase. This is not just about linting for syntactic sugar; Bandit dives deep into the code to unearth potential security flaws like SQL injection vulnerabilities, hardcoded credentials, insecure usage of cryptography, and even unsafe subprocess calls. The tool parses the Abstract Syntax Tree (AST) and conducts heuristic-based checks, offering vulnerability detection and an understanding of its context within the codebase. What elevates Bandit is its extensibility; you can write custom plugins using Python to meet specialized security needs, integrating seamlessly into CI/CD pipelines. In an era where DevSecOps is more than a buzzword, Bandit epitomizes how Python can actively contribute to making security an integral part of the development lifecycle.
Built on SaltStack, Hubble is a compliance-as-code framework that amplifies the capabilities of infrastructure-as-code (IAC) practices in enforcing security compliance. With Python at its core, Hubble allows on-demand auditing, real-time security notifications, and automatic remediation. Its modular architecture supports custom audit modules written in Python, thus enabling organizations to tailor their compliance rulesets. And let's pay attention to the real-time security notifications feature. Python's asynchronous capabilities, when leveraged within Hubble, can provide near-instantaneous alerting for any compliance drift, thereby enabling quicker remedial actions. Given the stringent regulatory landscapes organizations navigate today, Hubble's Python-driven extensibility is a lighthouse, guiding entities safely through the compliance storms.
Web applications remain one of the most prominent attack surfaces, and the backend often holds the keys to the kingdom. Secure.py is a need-to-know-and-have for backend web applications, particularly those developed using frameworks like Flask and Django. Secure.py is essentially a middleware that injects security headers into HTTP responses. Headers like Content Security Policy (CSP), X-Content-Type-Options, and Referrer-Policy are set up effortlessly, creating an added layer of security against common web-based vulnerabilities like XSS or data injection attacks. The library's Pythonic, decorator-based interface makes it trivial to integrate into existing Python web applications, thus enhancing security without disrupting the existing architecture.
The importance of cybersecurity
As the technology-centric wheel of modern civilization turns, the tapestry of our lives increasingly weaves into the digital sphere. This is not a phenomenon restricted to Silicon Valley or tech startups. The EU's digital transformation projects, banks' relentless push towards online services, and the private sector's stumbling transition to remote work are all a testament to the ever-growing reliance on the Internet and digital tools. Yet, this digital renaissance comes with a perilous underside – the escalating frequency and sophistication of cyber threats.
Western society is in a constant state of flux driven by technological innovation. The EU's strategic focus on digital transformation aims to catapult businesses and government services into a future where efficiency and accessibility are the norms. However, this transformation also exponentially expands the threat surface. We've seen this in the form of ill-prepared remote work setups, where endpoint security was hastily patched together or not implemented. Cybersecurity is often an afterthought in the rush to stay operational and adapt to new norms. But the stakes inevitably increase as we digitize more aspects of our lives and businesses.
Statistics on cyber threats often serve as mere abstractions, seldom inducing a sense of urgency. According to a recent report by ENISA, ransomware attacks nearly doubled in 2022, contributing to 38% of the threat landscape, followed by data-related threats at 30%, and a mix of DoS, DDoS, and RDoS attacks rounding out the scene. While these figures may seem academic to many, they represent the genuine battleground on which today's cyber warfare is fought.
Statistics: Faceless until you're one of them
There's often a disconnect in translating statistical data to personal vulnerability when it comes to risk assessment. Many businesses see these numbers as background noise until they become a casualty statistic. The inertia can partly be attributed to the "it won't happen to me" syndrome – an inherent underestimation of one's risk profile until one day, it does happen.
The Analog paradox: Cars, wallets, and cybersecurity
I often pose a question to clients and stakeholders: "Would you be comfortable leaving your car unlocked, wallet and house keys inside, along with personal information, unattended for months in the heart of a major city?" The unanimous answer is a resounding 'No.' Then why does cybersecurity often find itself at the bottom of priority lists? Unlike the tangibility of a car theft, the invisibility of cyber risks creates a cognitive dissonance. We often underestimate what we can't see or touch, and this complacency can be our downfall.
When your car grinds to a halt, you turn to your mechanic without hesitation. Their words become gospel when the doctor warns of an imminent health risk. Yet, when a cybersecurity professional sounds the alarm about potential risks that could compromise the very existence of a business, the reaction is often indifference or skepticism. The stakes are as real, if not more, but the sense of urgency is glaringly absent. This resistance to giving cybersecurity its due diligence has less to do with the discipline's legitimacy and more with the lack of widespread understanding of its critical importance.
As businesses and individuals navigate this digital world, viewing cybersecurity as a grudge purchase or a checkbox item can be a grievous error. The mesh of digital and physical worlds has left us in a state where cybersecurity isn't just an IT problem; it's a societal issue with ramifications that extend far beyond isolated incidents. The stakes are too high for it to be relegated to a footnote in any personal or corporate strategy.
Python is an educational catalyst
In my opinion, in the next few decades, knowing how to write code will become part of digital literacy. We should embrace that change. In a digital age where cybersecurity has become a critical societal concern, the schism between technical and non-technical talent poses a significant challenge. Python, a programming language celebrated for its readability and simplicity, emerges as an intriguing solution to this problem. Its structure, resembling pseudocode more than it does a labyrinth of brackets and semicolons, offers an inviting landscape for those daunted by the intricate syntax of languages like C++ or Java. This readability offers more than an entry point; it opens the gates to a domain often perceived as impenetrable.
The interactive nature of Python's shell enhances this sense of accessibility. Rather than confronting static documentation and abstract principles, users can interact in real-time with the code. They can run a command and see immediate results, transforming abstract concepts into tangible actions. This instant feedback demystifies the often-intimidating realm of cybersecurity, making it more digestible to novices. Even running a basic network scan or encrypting a file can dispel the aura of complexity surrounding cybersecurity tasks, fostering a sense of control and a curiosity to learn more.
The elegance of Python isn't confined to its structure or interactivity but extends to its practical application. One such application is in the visualization of complex cybersecurity data. Python's rich ecosystem includes a variety of data visualization libraries like Matplotlib and Seaborn. These tools can translate a threat report's abstract numbers and letters into relatable graphs and charts. When non-technical stakeholders can visually comprehend the stakes, the path to action becomes clearer, driving home the significance of cybersecurity in a language that transcends code.
The communal aspect of Python's ecosystem is another critical feature. An expansive array of resources, from tutorials to forums like Stack Overflow, empowers individuals to learn at their own pace and style. Companies can tap into this community-driven ethos by organizing internal workshops that bring together both technical and non-technical teams. These programs foster a shared sense of responsibility for cybersecurity, dispelling the notion that it's a subject only for the IT department.
However, the potential of Python as an educational tool isn't just in demystifying cybersecurity or expanding one's skill set. It has a psychological component that may be its most valuable asset – inspiration. There's an understated power in enabling someone to accomplish a task they previously thought beyond their abilities. When a non-technical team member can contribute to a cybersecurity solution with a few lines of Python code, it can ignite a passion for continuous learning. This inspiration is crucial, as cybersecurity isn't a static field but an evolving battleground that demands a workforce equipped with a lifelong learning mindset.
Furthermore, equipping non-technical talent with Python skills adds a layer of career versatility. These individuals now have the option to explore roles that intersect their primary skill sets with foundational cybersecurity knowledge. It's not just about creating a workforce of coders; it's about building a resilient, multifaceted team capable of navigating the ever-shifting labyrinth of today's cybersecurity landscape.
In essence, Python transcends its role as a programming language to become a catalyst for educational and cultural change within organizations. It challenges the status quo, bridges the technical and non-technical divide, and offers a holistic approach to cybersecurity. With Python as a mediator, cybersecurity becomes less of an exclusive domain for tech aficionados and more of an organization-wide competency, paving the way for a more secure and informed digital future.
Hiring is a breeze
In a corporate landscape where digital transformation is the norm, the roles of cybersecurity professionals have also evolved to become more complex and multi-faceted. The once-siloed domains of cybersecurity and development are increasingly overlapping, making professionals skilled in both realms a treasure trove for companies. Here's why recruiters find cybersecurity professionals with Python skills particularly alluring:
- Cost-effectiveness: Hiring separate teams for cybersecurity and Python development can be costly. Professionals adept at both can serve dual roles, thus reducing the overall cost of talent acquisition and retention.
- Quick turnaround: Python's syntactic simplicity allows cybersecurity experts to quickly script solutions to security challenges without getting bogged down by the complexities of the language. This short response time is invaluable in situations like incident response where every minute counts.
- Future-proofing: Cyber threats are evolving at a breakneck pace, and so should your defense mechanisms. Python's extensive libraries and community support enable continuous learning and adaptation, making it easier for companies to stay ahead of potential threats.
- Compliance and reporting: With regulations like GDPR and HIPAA requiring stringent data security measures, having team members capable of both implementing and validating security controls is a boon. Python skills are indispensable for developing custom scripts for compliance checks.
- Communication: Professionals who understand both cybersecurity and development can act as effective bridges between separate security and development teams, facilitating better communication and faster problem resolution.
- Innovation: With skills in Python, cybersecurity professionals can actively contribute to the development of proprietary security tools and solutions, adding unique value to the business.
- Reduced onboarding time: Cybersecurity professionals with Python skills can easily integrate into DevSecOps workflows, reducing the need for extensive onboarding or re-skilling.
- Culturally aligned: Professionals equipped with multidisciplinary skills like Python and cybersecurity are often naturally aligned with a culture of continuous improvement and agility – traits highly sought after in modern enterprises.
The talent pool and the upskilling of Python developers into cybersecurity roles is also not too difficult. There are a few factors involved in why that is the case:
- Abundant talent pool: Python is one of the most popular languages, which means there's a large talent pool to tap into. While not all Python developers will have a cybersecurity background, their existing Python knowledge offers a head start for upskilling into cybersecurity roles.
- Low barrier to entry: For Python developers, the transition into cybersecurity roles is less strenuous. Familiarity with syntax and coding constructs allows them to easily pick up security-specific Python libraries and frameworks, making them productive in a shorter time frame.
- Online resources and communities: Python developers looking to transition into cybersecurity have various resources at their disposal. From dedicated forums to specialized training programs and certifications such as OSCP, CEH, or CompTIA Security+, upskilling can be self-paced or structured.
- Hands-on learning: Python developers already have the coding prowess to experiment and implement. With tools like Metasploit Unleashed or WebGoat available for hands-on learning, transitioning into roles like penetration testing becomes more straightforward.
- Maturity in cybersecurity career: As Python developers delve deeper into cybersecurity, they can specialize further. Whether cloud security, endpoint protection, or threat hunting, Python skills remain a common underpinning, enabling career growth and specialization.
- Business acumen: Python developers often possess a knack for solving real-world problems through code. When this practical approach is applied to cybersecurity, it results in pragmatic, business-friendly solutions that align security with organizational objectives.
- AI and ML in cybersecurity: Python is integral in machine learning and AI – technologies that are becoming essential in predictive cybersecurity. Skilled Python developers can readily contribute to initiatives involving anomaly detection, threat prediction, and automated incident response.
- Consulting and leadership roles: For those Python developers who move into cybersecurity and garner substantial experience, avenues open up for consultancy and leadership roles, which demand a blend of technical and business acumen.
Integrating Python skills into cybersecurity roles offers a unique blend of capabilities that streamline operations, enhance security, and add strategic value to the organization. Conversely, Python developers find a smoother and more diversified career trajectory when they branch into cybersecurity, backed by the foundational skills they already possess. It's a win-win situation that HR and talent management teams would do well to capitalize on.
Summary
In a rapidly digitizing world where cybersecurity risks continue to increase, Python emerges as an educational catalyst, not just as a tool. Its intuitive syntax and interactive nature make it an accessible medium for bridging the divide between technical and non-technical talent within organizations. Python's rich ecosystem, from its libraries to its vibrant community, further offers a pathway for collective learning. The language transcends its technical applications to foster a culture of continuous learning and shared responsibility for cybersecurity. By embracing Python as a medium for education, organizations are better equipped to navigate the complex and ever-evolving landscape of cybersecurity risks. With Python at the helm, cybersecurity is transformed from a specialist's burden to an organization-wide competency, fortifying our collective defense against future digital threats.
