<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Andreyna Marques Carvalho on Medium]]></title>
        <description><![CDATA[Stories by Andreyna Marques Carvalho on Medium]]></description>
        <link>https://medium.com/@andreyna.m.carvalho?source=rss-a477eba5584e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*B4Jt4_auMYArtHYzu4iBpQ.jpeg</url>
            <title>Stories by Andreyna Marques Carvalho on Medium</title>
            <link>https://medium.com/@andreyna.m.carvalho?source=rss-a477eba5584e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 08:37:11 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@andreyna.m.carvalho/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[OSI Model: My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/osi-model-my-beginner-friendly-notes-cd577a08c7dc?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/cd577a08c7dc</guid>
            <category><![CDATA[osi-model]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Mon, 06 Jul 2026 19:15:31 GMT</pubDate>
            <atom:updated>2026-07-06T19:15:31.934Z</atom:updated>
            <content:encoded><![CDATA[<p>When I started studying networking and cybersecurity, one topic appeared again and again: the <strong>OSI Model</strong>.</p><p>At first, it can look like just another theory with seven layers to memorize. But after studying it more carefully, I started to understand that the OSI Model is actually a map that helps explain how data travels from one device to another.</p><p>Instead of looking at network communication as one big complicated process, the OSI Model breaks it into smaller parts.</p><p>Each layer has a specific responsibility. Some layers deal with cables and signals. Others deal with MAC addresses, IP addresses, connections, encryption, and applications.</p><p>In this article, I want to share my beginner-friendly notes about the OSI Model and how I understood each layer.</p><h3>About the OSI Model</h3><p>The <strong>OSI Model</strong> stands for <strong>Open Systems Interconnection Model</strong>.</p><p>In simple words, it is a framework that helps explain how devices communicate over a network.</p><p>The model has <strong>seven layers</strong>, and each layer plays a different role in the communication process.</p><p>The idea that helped me understand it better was this:</p><p>When data travels across a network, it does not magically reach the destination. It passes through different stages, and each stage adds or handles specific information.</p><p>This process is called <strong>encapsulation</strong>.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Open Systems Interconnection</em></blockquote><blockquote><em>7</em></blockquote><blockquote><em>encapsulation</em></blockquote><h3>About Layer 1: Physical Layer</h3><p>The <strong>Physical Layer</strong> is the easiest layer to visualize because it deals with the real physical parts of the network.</p><p>Here we are talking about cables, electrical signals, network hardware, and the transmission of raw bits.</p><p>A simple way I understood this layer is:</p><p>If the network were a road, the Physical Layer would be the actual road where the traffic moves.</p><p>In networking, the “traffic” is made of bits, represented as <strong>1s and 0s</strong>. These bits move through physical media such as Ethernet cables, fiber cables, or wireless signals.</p><p>So, when we talk about the Physical Layer, we are not thinking about websites, browsers, or IP addresses yet. We are thinking about the basic physical movement of data.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Physical</em></blockquote><blockquote><em>Binary</em></blockquote><blockquote><em>Ethernet Cables</em></blockquote><h3>About Layer 2: Data Link Layer</h3><p>The <strong>Data Link Layer</strong> focuses on physical addressing.</p><p>This is where <strong>MAC addresses</strong> become important.</p><p>Every network-enabled device has a piece of hardware called a <strong>Network Interface Card</strong>, also known as a NIC. This card has a unique MAC address, which helps identify the device inside a local network.</p><p>The way I like to think about it is:</p><p>An IP address is like the logical address of a device, while the MAC address is like the physical identity of the network card.</p><p>This layer helps make sure data is prepared correctly and sent to the right device inside the local network.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Data Link</em></blockquote><blockquote><em>Network Interface Card</em></blockquote><h3>About Layer 3: Network Layer</h3><p>The <strong>Network Layer</strong> is where routing starts to happen.</p><p>This layer works with <strong>IP addresses</strong> and is responsible for deciding the best path for data to travel from one network to another.</p><p>This is where routers become important.</p><p>A router is considered a Layer 3 device because it works with IP addresses and helps move packets between different networks.</p><p>A simple way to understand this layer is:</p><p>If the Data Link Layer helps devices communicate inside the same local network, the Network Layer helps data travel between different networks.</p><p>This layer also introduces routing protocols such as <strong>OSPF</strong> and <strong>RIP</strong>, which help routers decide the best path for traffic.</p><p>Some factors that can influence the route are the shortest path, the most reliable path, and the fastest physical connection.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Network</em></blockquote><blockquote><em>Y</em></blockquote><blockquote><em>Open Shortest Path First</em></blockquote><blockquote><em>Routing Information Protocol</em></blockquote><blockquote><em>IP Addresses</em></blockquote><h3>About Layer 4: Transport Layer</h3><p>The <strong>Transport Layer</strong> is one of the most important layers to understand because it deals with how data is delivered.</p><p>This layer introduces two famous protocols: <strong>TCP</strong> and <strong>UDP</strong>.</p><p>TCP stands for <strong>Transmission Control Protocol</strong>.</p><p>TCP is focused on reliability. It checks if the data arrived correctly, keeps things in order, and makes sure the receiver gets the full message.</p><p>I understood TCP like this:</p><p>TCP is like sending an important document and asking for confirmation that every page arrived correctly.</p><p>That is why TCP is commonly used for things like web browsing, emails, and file transfers.</p><p>UDP stands for <strong>User Datagram Protocol</strong>.</p><p>UDP is faster, but it does not guarantee delivery. It sends the data without worrying too much about whether every piece arrived perfectly.</p><p>I understood UDP like this:</p><p>UDP is like sending information quickly without waiting for confirmation.</p><p>That is useful for things like video streaming, voice calls, and gaming, where speed can be more important than perfect delivery.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Transport</em></blockquote><blockquote><em>Transmission Control Protocol</em></blockquote><blockquote><em>User Datagram Protocol</em></blockquote><blockquote><em>TCP</em></blockquote><blockquote><em>UDP</em></blockquote><blockquote><em>TCP</em></blockquote><blockquote><em>TCP</em></blockquote><blockquote><em>UDP</em></blockquote><h3>About Layer 5: Session Layer</h3><p>The <strong>Session Layer</strong> is responsible for creating, maintaining, and closing sessions between devices.</p><p>A session begins when a connection is successfully established.</p><p>This layer makes sure that communication between two devices stays organized while the connection is active.</p><p>One detail I found interesting is that sessions can also have checkpoints. This means that if something goes wrong, the communication does not always need to start again from zero.</p><p>A simple way to understand the Session Layer is:</p><p>It manages the conversation between two devices.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Session</em></blockquote><blockquote><em>Session</em></blockquote><h3>About Layer 6: Presentation Layer</h3><p>The <strong>Presentation Layer</strong> acts like a translator.</p><p>Its job is to make sure data is formatted in a way that the receiving system can understand.</p><p>This matters because different applications can work in different ways, but the data still needs to make sense on both sides.</p><p>A good example is email. One person may use one email client, and another person may use a different one, but the message still needs to appear correctly.</p><p>This layer is also related to things like encryption, including secure communication such as HTTPS.</p><p>The way I understood this layer is:</p><p>The Presentation Layer prepares the data so the application can understand it properly.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Presentation</em></blockquote><blockquote><em>Translator</em></blockquote><h3>About Layer 7: Application Layer</h3><p>The <strong>Application Layer</strong> is the closest layer to the user.</p><p>This is the layer we interact with more directly when using browsers, email clients, file transfer tools, and other applications.</p><p>It does not mean the application itself is the layer. Instead, this is where protocols and rules exist so applications can communicate over the network.</p><p>For example, when we access a website, use DNS, send an email, or transfer a file, we are working with concepts related to the Application Layer.</p><p>A simple way to remember this layer is:</p><p>The Application Layer is where the user experience meets the network.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Application</em></blockquote><blockquote><em>Graphical User Interface</em></blockquote><h3>About the OSI Review Challenge</h3><p>At the end of the study, there was an interactive review challenge focused on remembering the OSI layers in the correct order.</p><p>This was useful because instead of only reading theory, I had to apply the sequence and reinforce the structure of the model.</p><p>The OSI layers from Layer 1 to Layer 7 are:</p><p>Physical<br> Data Link<br> Network<br> Transport<br> Session<br> Presentation<br> Application</p><p>For public notes, I prefer replacing the original platform acronym with a generic placeholder while keeping the structure understandable.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>ACRONYM_HERE{OSI_DUNGEON_ESCAPED}</em></blockquote><h3>What I Learned</h3><p>This study helped me understand the OSI Model in a much clearer way.</p><p>Before studying it, the OSI Model can feel like something we only need to memorize for exams. But after breaking it down layer by layer, it starts to make more sense.</p><p>Each layer has a purpose.</p><p>The <strong>Physical Layer</strong> moves raw bits.</p><p>The <strong>Data Link Layer</strong> uses MAC addresses.</p><p>The <strong>Network Layer</strong> uses IP addresses and routing.</p><p>The <strong>Transport Layer</strong> handles TCP and UDP.</p><p>The <strong>Session Layer</strong> manages connections.</p><p>The <strong>Presentation Layer</strong> translates and formats data.</p><p>The <strong>Application Layer</strong> supports user-facing network services.</p><p>For anyone starting in cybersecurity, understanding the OSI Model is very important. It helps when studying networking, troubleshooting, firewalls, packet analysis, SOC alerts, and even incident response.</p><p>The better we understand how data moves, the better we can understand how to protect it.</p><h3>Final Thoughts</h3><p>Learning the OSI Model was a great step in my networking fundamentals journey.</p><p>It helped me review the theory while also connecting each layer with practical examples.</p><p>As a beginner, I am still building my foundation, but this topic helped me understand network communication with more structure.</p><p>This was another step in my cybersecurity learning journey.</p><p>Let’s evolve together.</p><h3>tags</h3><p>Cybersecurity, OSI Model, Networking, Network Security, Blue Team, SOC, Information Security, TCP, UDP, Data Link, Network Layer, Cybersecurity Beginner, Learning Notes</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/713/1*e8V6GD0wyXF_Mq4fMABJaA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cd577a08c7dc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Understanding LAN Fundamentals: My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/intro-to-lan-tryhackme-completed-def23bf86840?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/def23bf86840</guid>
            <category><![CDATA[networking]]></category>
            <category><![CDATA[intro-to-lan]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Wed, 01 Jul 2026 18:21:25 GMT</pubDate>
            <atom:updated>2026-07-03T19:28:07.412Z</atom:updated>
            <content:encoded><![CDATA[<p>When I started studying networking, I realized that some concepts appear again and again in cybersecurity.</p><p>LANs, subnetting, ARP, DHCP, switches, routers, gateways — at first, they may look like separate topics. But after studying them together, everything starts to connect.</p><p>This article is a simple summary of what I learned about local networks and how devices communicate inside them.</p><h3>About LAN Topologies</h3><p>A <strong>Local Area Network</strong> is one of the most common types of networks we use every day.</p><p>When a computer connects to a printer in the same office, when a laptop accesses a local server, or when multiple devices share the same internet connection inside a company, they are usually part of a LAN.</p><p>Inside this environment, routers and switches play very important roles.</p><p>A router performs <strong>routing</strong>, which means sending data from one network to another. A switch, on the other hand, connects multiple devices inside the same local network and helps send data to the correct destination.</p><p>Network topology is also important because it defines how devices are connected. A <strong>Bus Topology</strong> is usually cheaper to set up, while a <strong>Star Topology</strong> is more expensive but easier to manage and troubleshoot.</p><p>In simple words, every topology has trade-offs. Some are cheaper, some are easier to maintain, and some are more reliable.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Local Area Network</blockquote><blockquote>Routing</blockquote><blockquote>Switch</blockquote><blockquote>Bus Topology</blockquote><blockquote>Star Topology</blockquote><blockquote>… Just you…</blockquote><h3>About Subnetting</h3><p>Subnetting is one of those topics that looks scary at first, but it becomes easier when we understand the purpose behind it.</p><p>In simple words, <strong>subnetting</strong> means dividing a network into smaller parts.</p><p>This helps organize devices, reduce unnecessary traffic, and make the network easier to manage. Instead of having one large network with everything mixed together, subnetting allows us to separate it into smaller logical sections.</p><p>In IPv4, a subnet mask has <strong>32 bits</strong>. Each section of an IPv4 address is called an octet, and each octet can range from <strong>0 to 255</strong>.</p><p>Another important point is understanding the difference between a network address and a host address.</p><p>The <strong>network address</strong> identifies the start of a network, while the <strong>host address</strong> identifies devices inside that network.</p><p>And when a device needs to communicate outside its own network, it usually sends the traffic to the <strong>default gateway</strong>.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Subnetting</blockquote><blockquote>32</blockquote><blockquote>0–255</blockquote><blockquote>Network Address</blockquote><blockquote>Host Address</blockquote><blockquote>Default Gateway</blockquote><h3>About ARP</h3><p>ARP is a very important protocol for local network communication.</p><p>It stands for <strong>Address Resolution Protocol</strong>.</p><p>The idea behind ARP is simple: sometimes a device knows the IP address it wants to communicate with, but it still needs to discover the MAC address of that device.</p><p>A simple way to imagine ARP is like a device asking the local network:</p><p>“Who has this IP address?”</p><p>That message is an <strong>ARP Request</strong>.</p><p>This matters because an <strong>IP Address</strong> is a logical identifier, while a <strong>MAC Address</strong> is a physical identifier used by devices inside the local network.</p><p>So, ARP helps connect these two worlds: logical addressing and physical addressing.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Address Resolution Protocol</blockquote><blockquote>Request</blockquote><blockquote>MAC Address</blockquote><blockquote>IP Address</blockquote><h3>About DHCP</h3><p>DHCP is another protocol that makes networking much easier in real life.</p><p>It stands for <strong>Dynamic Host Configuration Protocol</strong>.</p><p>Without DHCP, we would need to manually configure IP addresses, subnet masks, gateways, and DNS servers on every device. That would take a lot of time, especially in companies or larger networks.</p><p>DHCP automates this process.</p><p>When a device connects to a network and needs an IP address, it starts by sending a <strong>DHCP Discover</strong> packet.</p><p>After the server offers an IP address, the device responds with a <strong>DHCP Request</strong>.</p><p>Finally, the server confirms everything with a <strong>DHCP ACK</strong>.</p><p>A simple way to remember the process is:</p><p>Discover → Offer → Request → ACK</p><p>This is happening in the background almost every time we connect a phone, laptop, or computer to a network.</p><p><strong>My learning notes from this section:</strong></p><blockquote>DHCP Discover</blockquote><blockquote>DHCP Request</blockquote><blockquote>DHCP ACK</blockquote><h3>What I Learned</h3><p>This study helped me understand that networking is not just about memorizing terms.</p><p>A LAN connects devices in a local environment.</p><p>A router sends data between networks.</p><p>A switch connects devices inside the same local network.</p><p>Subnetting divides networks into smaller parts.</p><p>ARP helps devices discover MAC addresses.</p><p>DHCP automatically gives network configuration to devices.</p><p>These fundamentals are extremely important for cybersecurity because before we can understand attacks, alerts, firewalls, SIEM logs, or packet analysis, we need to understand how normal network communication works.</p><h3>Final Thoughts</h3><p>Learning networking fundamentals is one of the first steps in building a strong cybersecurity foundation.</p><p>Topics like LANs, subnetting, ARP, DHCP, switches, routers, and gateways may look basic, but they are used everywhere in real environments.</p><p>The better we understand how networks work, the better we can understand how to protect them.</p><p>Let’s evolve together.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=def23bf86840" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Is Networking? My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/what-is-networking-tryhackme-11d00640de0f?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/11d00640de0f</guid>
            <category><![CDATA[what-is-network]]></category>
            <category><![CDATA[network-security]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[networking]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Mon, 29 Jun 2026 20:52:53 GMT</pubDate>
            <atom:updated>2026-07-03T19:54:44.287Z</atom:updated>
            <content:encoded><![CDATA[<p>When I started studying networking, I realized that almost everything in cybersecurity depends on understanding how devices communicate.</p><p>Before learning about attacks, alerts, firewalls, SIEM logs, or incident response, we need to understand the basics: what a network is, how the internet works, how devices are identified, and how we can test connectivity.</p><p>This article is a simple summary of my beginner-friendly notes about networking fundamentals.</p><h3>About Networking</h3><p>A network is created when devices are connected together.</p><p>This can be something simple, like two computers sharing files, or something larger, like a company network with computers, servers, printers, switches, routers, and wireless devices.</p><p>The important idea is that a network allows devices to communicate and exchange data.</p><p>In cybersecurity, this concept is extremely important because attackers and defenders both interact with networks. If we understand how devices communicate normally, it becomes easier to identify suspicious behavior later.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Network</blockquote><h3>About the Internet</h3><p>The internet is basically a huge network made of many smaller networks connected together.</p><p>When we access websites, send emails, watch videos, or use online services, our devices are communicating through this global network.</p><p>One important historical point is the creation of the World Wide Web.</p><p>The World Wide Web was invented by Tim Berners-Lee. It helped make the internet more accessible by allowing people to browse websites through links and web pages.</p><p>A simple way to understand the difference is this:</p><p>The internet is the infrastructure.</p><p>The World Wide Web is one of the services that runs on top of it.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Tim Berners-Lee</blockquote><h3>About Identifying Devices on a Network</h3><p>Every device on a network needs a way to be identified.</p><p>One of the most common identifiers is the IP address.</p><p>IP stands for Internet Protocol. It is used to identify devices and help data reach the correct destination across networks.</p><p>In IPv4, an IP address is divided into four sections. Each section is called an octet.</p><p>For example, an IPv4 address usually looks like this:</p><p>192.168.1.10</p><p>Each part separated by dots is an octet, and IPv4 addresses have four sections.</p><p>Another important identifier is the MAC address.</p><p>MAC stands for Media Access Control. Unlike an IP address, which is logical and can change depending on the network configuration, a MAC address is related to the physical network interface of the device.</p><p>A simple way to remember it is:</p><p>An IP address helps identify where a device is on a network.</p><p>A MAC address helps identify the physical network interface of the device.</p><p>Both are important for communication, especially inside local networks.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Internet Protocol</blockquote><blockquote>Octet</blockquote><blockquote>4</blockquote><blockquote>Media Access Control</blockquote><h3>About Ping and ICMP</h3><p>Ping is one of the simplest and most useful tools for testing network connectivity.</p><p>When we use ping, we are checking whether another device or server can be reached over the network.</p><p>Ping uses ICMP, which stands for Internet Control Message Protocol.</p><p>A basic ping command follows a very simple structure. For example, to test connectivity with the IP address 10.10.10.10, the command would be:</p><p>ping 10.10.10.10</p><p>This command sends ICMP packets to the target and waits for a response.</p><p>If the target responds, it usually means there is network connectivity between your device and that destination. If it does not respond, there may be a network issue, firewall rule, routing problem, or the target may simply be configured not to reply.</p><p>In real environments, ping is often one of the first tests used during troubleshooting.</p><p><strong>My learning notes from this section:</strong></p><blockquote>ICMP</blockquote><blockquote>ping 10.10.10.10</blockquote><h3>Continue Learning: LAN Fundamentals</h3><p>After understanding basic networking concepts, the next natural step is learning about LANs.</p><p>LAN stands for Local Area Network, and it helps explain how devices communicate inside a local environment, such as a home, office, or company.</p><p>Studying LAN fundamentals also makes it easier to understand switches, routers, subnetting, ARP, DHCP, and the OSI Model.</p><p>My learning notes from this section:</p><p>No answer needed</p><h3>What I Learned</h3><p>This study helped me understand that networking starts with simple but powerful concepts.</p><p>A network is formed when devices are connected together.</p><p>The World Wide Web was invented by Tim Berners-Lee.</p><p>IP means Internet Protocol.</p><p>Each section of an IPv4 address is called an octet.</p><p>An IPv4 address has four sections.</p><p>MAC means Media Access Control.</p><p>Ping uses ICMP to test connectivity.</p><p>The ping command is one of the most basic tools for checking whether a device can be reached.</p><p>These fundamentals may look simple, but they are used all the time in cybersecurity, infrastructure, support, SOC analysis, and incident response.</p><p>Before we can understand attacks, we need to understand normal communication.</p><h3>Final Thoughts</h3><p>Learning networking fundamentals is one of the best first steps for anyone starting in cybersecurity.</p><p>Concepts like networks, IP addresses, MAC addresses, ICMP, and ping appear constantly in real-world environments.</p><p>The better we understand how devices communicate, the better we can understand how to protect them.</p><p>Let’s evolve together.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=11d00640de0f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Careers in Cyber: My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/tryhackme-careers-in-cyber-completed-walkthrough-e86fc7eff062?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/e86fc7eff062</guid>
            <category><![CDATA[information-security]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[careers-in-cyber]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Thu, 25 Jun 2026 11:26:16 GMT</pubDate>
            <atom:updated>2026-07-03T20:00:38.590Z</atom:updated>
            <content:encoded><![CDATA[<p>When I started learning more about cybersecurity, one thing became clear very quickly: this field is much bigger than most people imagine.</p><p>Cybersecurity is not only about hackers, passwords, or technical tools. There are many different career paths, and each one has a different role in protecting systems, networks, data, and people.</p><p>Some professionals focus on monitoring alerts. Others build security solutions. Some test systems for weaknesses. And some help organizations improve their processes, policies, and defenses.</p><p>This article is a simple summary of my notes about different careers in cyber and what I understood from each role.</p><h3>About Cybersecurity Careers</h3><p>Cybersecurity is a growing field with a huge demand for professionals.</p><p>One number that caught my attention was <strong>3.5 million</strong> unfilled cyber positions. This shows how much companies need people who understand security and can help protect digital environments.</p><p>But what I found interesting is that cybersecurity is not a single career path.</p><p>There are many possibilities, such as security analysis, security engineering, penetration testing, incident response, governance, threat intelligence, cloud security, and many others.</p><p>In simple words, cybersecurity has space for different profiles. Some people enjoy investigation. Some enjoy building systems. Some enjoy breaking things ethically to help improve security. Others enjoy documentation, risk management, and strategy.</p><p><strong>My learning notes from this section:</strong></p><blockquote>3.5 million</blockquote><h3>About Security Analysis</h3><p>Security analysts play a very important role in an organization’s <strong>defence</strong>.</p><p>This role is usually connected to monitoring, investigating alerts, analyzing suspicious activity, and helping the organization respond to possible threats.</p><p>A security analyst may work with SIEM tools, logs, dashboards, alerts, tickets, and incident response processes.</p><p>I like to think of this role as one of the “eyes” of the security team.</p><p>The analyst watches what is happening, investigates when something looks strange, and helps the organization understand whether an event is normal or suspicious.</p><p>For someone starting in cybersecurity, this is one of the most interesting paths because it builds a strong foundation in real-world security operations.</p><p><strong>My learning notes from this section:</strong></p><blockquote>defence</blockquote><h3>About Security Engineering</h3><p>Security engineering is more focused on building, maintaining, and improving security systems.</p><p>While a security analyst may spend more time investigating alerts, a security engineer often works on the tools and infrastructure that help protect the environment.</p><p>One example of a device or system related to this role is an <strong>Intrusion Detection System</strong>.</p><p>An Intrusion Detection System helps identify suspicious activity in a network or system. It can detect signs of attacks, unusual behavior, or traffic that does not look normal.</p><p>In simple words, if the security analyst is watching the alerts, the security engineer helps make sure the security tools are working properly.</p><p>This role requires strong technical knowledge because it often involves networks, systems, security architecture, configurations, and tool maintenance.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Intrusion Detection System</blockquote><h3>About Penetration Testing</h3><p>Penetration testing is one of the most well-known areas in cybersecurity.</p><p>A penetration tester looks for weaknesses in systems, applications, networks, or organizations before malicious attackers can exploit them.</p><p>The process followed during this type of work is called an <strong>engagement</strong>.</p><p>An engagement usually has rules, scope, objectives, and permission. This is very important because ethical hacking must be controlled, authorized, and professional.</p><p>In simple words, penetration testing is not about “hacking anything.” It is about testing security in a responsible way to help the organization improve.</p><p>This career path is great for people who enjoy problem-solving, technical challenges, research, and offensive security.</p><p><strong>My learning notes from this section:</strong></p><blockquote>engagement</blockquote><h3>About Choosing a Cybersecurity Path</h3><p>One thing I liked about studying cybersecurity careers is that it made me think about which role fits me the most.</p><p>There is no single “best” career in cyber. The best path depends on your interests, skills, and the kind of work that motivates you.</p><p>Some people may prefer Blue Team roles, such as SOC analyst, incident responder, or security analyst.</p><p>Others may prefer Red Team roles, such as penetration tester or red team operator.</p><p>There are also people who enjoy governance, risk, compliance, cloud security, security awareness, digital forensics, malware analysis, or security engineering.</p><p>In my case, the role that caught my attention was <strong>Red Teamer</strong>.</p><p>That does not mean I only want to learn offensive security, but it shows that I enjoy understanding how attacks work so I can also understand how to defend against them better.</p><p><strong>My learning notes from this section:</strong></p><blockquote>I’ve identified my dream cyber security role!</blockquote><blockquote>Red Teamer</blockquote><h3>What I Learned</h3><p>This study helped me understand that cybersecurity is not just one job.</p><p>There are many roles, and each one contributes to security in a different way.</p><p>A <strong>Security Analyst</strong> helps defend the organization by monitoring, investigating, and responding to threats.</p><p>A <strong>Security Engineer</strong> helps build and maintain security systems, such as an Intrusion Detection System.</p><p>A <strong>Penetration Tester</strong> follows an engagement process to find weaknesses before attackers do.</p><p>A <strong>Red Teamer</strong> thinks like an attacker to help improve the organization’s defenses.</p><p>The biggest lesson for me is that cybersecurity has many doors. You do not need to know everything at the beginning. You need to start learning, explore different areas, and slowly understand which path makes more sense for you.</p><h3>Final Thoughts</h3><p>Learning about cybersecurity careers helped me see the field with a broader perspective.</p><p>Cybersecurity is not only technical. It also involves strategy, communication, documentation, analysis, creativity, and continuous learning.</p><p>For anyone starting in this area, my advice is simple: explore.</p><p>Try to understand the different roles, learn the fundamentals, and pay attention to what makes you curious.</p><p>Every step helps you get closer to your path.</p><p>Let’s evolve together.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e86fc7eff062" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Defensive Security Intro: My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/tryhackme-defensive-security-intro-completed-3618f631b3ec?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/3618f631b3ec</guid>
            <category><![CDATA[defensive-security]]></category>
            <category><![CDATA[defensive-security-intro]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Wed, 24 Jun 2026 11:18:56 GMT</pubDate>
            <atom:updated>2026-07-03T20:11:42.659Z</atom:updated>
            <content:encoded><![CDATA[<p>When I started learning cybersecurity, I noticed that many people first think about hacking, exploitation, and offensive techniques.</p><p>But cybersecurity is not only about attacking systems.</p><p>A huge part of this field is about <strong>defensive security</strong>: detecting suspicious activity, investigating what happened, responding to incidents, and helping protect the environment before the damage becomes bigger.</p><p>This article is a simple summary of my beginner-friendly notes about defensive security and what I learned from analyzing suspicious activity in a small practice scenario.</p><h3>About Defensive Security</h3><p>Defensive security focuses on protecting systems, networks, users, and data.</p><p>The main goal is to <strong>detect and respond to attacks</strong>.</p><p>In simple words, defensive security is about watching what is happening, identifying suspicious behavior, understanding the possible threat, and taking action to reduce the impact.</p><p>This is very connected to Blue Team work, SOC analysis, monitoring, incident response, log analysis, and threat detection.</p><p>A good defender needs to understand normal behavior first. Only then it becomes easier to notice when something does not look right.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Detect and respond to attacks</blockquote><h3>About Detecting Suspicious Activity</h3><p>One of the first steps in defensive security is learning how to identify suspicious activity.</p><p>In a real environment, this could involve checking logs, alerts, firewall events, authentication attempts, web requests, or network traffic.</p><p>During this kind of analysis, one important detail is the <strong>source IP address</strong>.</p><p>The source IP can help us understand where the suspicious traffic is coming from. It does not always tell the full story, but it gives us a starting point for investigation.</p><p>For example, if many unusual requests are coming from the same IP address, that may indicate scanning, brute force attempts, automated activity, or another suspicious behavior.</p><p>The key lesson here is that defenders need to pay attention to patterns.</p><p>One request may not mean much. But repeated suspicious activity from the same source can tell a different story.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Source IP address</blockquote><blockquote>Suspicious traffic</blockquote><blockquote>Traffic analysis</blockquote><blockquote>e.g. 32.122.195.63</blockquote><h3>About Identifying the Attack</h3><p>After detecting suspicious activity, the next step is to understand what the attacker is trying to do.</p><p>In web environments, this often means looking at requested URLs, paths, parameters, and HTTP activity.</p><p>A very common behavior is when an attacker tries to find sensitive or restricted pages, such as administrative panels, login pages, backup files, or hidden directories.</p><p>For example, when someone tries to access an admin area without permission, that can be a sign of reconnaissance or an attempt to compromise the application.</p><p>This part helped me understand that defensive security is not just about seeing an alert and stopping there.</p><p>We need to ask questions like:</p><p>What was accessed?</p><p>Where did the request come from?</p><p>Was the request successful?</p><p>Is the behavior normal?</p><p>Could this be part of a bigger attack?</p><p><strong>My learning notes from this section:</strong></p><blockquote>URL analysis</blockquote><blockquote>Admin page attempt</blockquote><blockquote>Attack identification</blockquote><blockquote>e.g. <a href="https://fakebank.com/admin">https://fakebank.com/admin</a></blockquote><h3>About Stopping the Attack</h3><p>Once suspicious activity is detected and the attack is understood, the next step is response.</p><p>Stopping an attack may involve blocking an IP address, disabling a compromised account, updating firewall rules, isolating a machine, removing malicious files, or fixing a vulnerable configuration.</p><p>The correct response depends on the situation.</p><p>In a simple training scenario, stopping the attack may look straightforward. But in real life, defenders need to be careful and document everything.</p><p>A rushed response can sometimes break services, remove evidence, or hide important details that would help the investigation.</p><p>That is why incident response is not only technical. It also requires process, communication, and good decision-making.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Containment</blockquote><blockquote>Response</blockquote><blockquote>Attack stopped</blockquote><h3>Why Defensive Security Matters</h3><p>Defensive security is important because organizations are constantly exposed to threats.</p><p>Attackers may try to steal data, compromise accounts, exploit vulnerabilities, disrupt services, or move laterally inside a network.</p><p>Blue Team professionals help reduce this risk by monitoring environments and responding when something suspicious happens.</p><p>Some common defensive activities include:</p><p>Monitoring alerts.</p><p>Analyzing logs.</p><p>Investigating suspicious IP addresses.</p><p>Reviewing web requests.</p><p>Identifying attack patterns.</p><p>Responding to incidents.</p><p>Improving security controls.</p><p>This is one of the reasons why defensive security is such an important area for beginners to understand.</p><p>It builds the foundation for roles like SOC Analyst, Security Analyst, Incident Responder, Threat Hunter, and Blue Team Analyst.</p><h3>What I Learned</h3><p>This study helped me understand that defensive security is about much more than tools.</p><p>The main goal is to <strong>detect and respond to attacks</strong>.</p><p>A suspicious source IP can be an important clue during an investigation.</p><p>Requested URLs can reveal what the attacker is trying to access.</p><p>Identifying the attack is only part of the process.</p><p>A good defender also needs to respond, contain the threat, and help prevent similar incidents in the future.</p><p>The biggest lesson for me is that cybersecurity defense starts with curiosity.</p><p>When something looks suspicious, we need to investigate, understand the context, and make decisions based on evidence.</p><h3>Final Thoughts</h3><p>Learning defensive security gave me a better understanding of the Blue Team side of cybersecurity.</p><p>It showed me that protecting systems requires attention, analysis, and continuous learning.</p><p>As a beginner, I am still building my foundation, but each topic helps me understand the field a little better.</p><p>I am also using this journey to improve my English while learning cybersecurity, and I hope these notes can help other beginners who are starting too.</p><p>If you are also learning cybersecurity or English, feel free to follow this journey with me.</p><p>Let’s evolve together.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3618f631b3ec" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[TryHackme — Offensive Security Intro [COMPLETED]]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/tryhackme-offensive-security-intro-completed-36567bfa8c8d?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/36567bfa8c8d</guid>
            <category><![CDATA[offensive-security-intro]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Sun, 21 Jun 2026 15:14:31 GMT</pubDate>
            <atom:updated>2026-07-03T21:04:44.855Z</atom:updated>
            <content:encoded><![CDATA[<h3>Offensive Security Intro: My Beginner-Friendly Notes</h3><p>For educational purposes only.</p><p>When I started learning cybersecurity, one of the first things I noticed is that offensive security is often misunderstood.</p><p>Some people think it is only about “hacking,” but in a professional and ethical context, offensive security is about simulating attacker behavior to help identify weaknesses before real attackers can exploit them.</p><p>This article is a simple summary of my beginner-friendly notes from my first steps into offensive security.</p><h3>About Thinking Like a Hacker</h3><p>Offensive security is about thinking like an attacker, but with permission and a clear purpose.</p><p>The goal is not to cause damage. The goal is to understand how systems can be attacked so we can make them stronger.</p><p>In simple words, offensive security helps us find vulnerabilities before malicious hackers do.</p><p>It is one of the first concepts that helped me understand the difference between illegal hacking and ethical hacking.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>Offensive Security</em></blockquote><h3>About Starting the Lab</h3><p>The practical part introduced a very basic web application scenario.</p><p>The idea was simple: interact with the application, observe the information displayed, and start understanding how a vulnerable environment can be tested.</p><p>Even though the example was very beginner-friendly, it helped me understand that offensive security often starts with observation.</p><p>Before attacking anything, we need to understand what the application shows, how it behaves, and what information is available.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>8881</em></blockquote><h3>About Finding Hidden Pages</h3><p>Another important concept was discovering hidden pages.</p><p>Sometimes, web applications have pages that are not directly visible through the main interface. These pages may still exist on the server and can sometimes be discovered through directory enumeration.</p><p>This is important because hidden does not always mean protected.</p><p>If a sensitive page exists but has no proper access control, it can become a security risk.</p><p>In this example, the hidden page represented a part of the application that should not be exposed without proper protection.</p><p><strong>My learning notes from this section:</strong></p><blockquote><a href="http://fakebank.thm/bank-transfer"><em>http://fakebank.thm/bank-transfer</em></a></blockquote><h3>About Attacking the Admin Page</h3><p>The final part showed a very simple example of how a vulnerable page could be abused.</p><p>The goal was not to learn an advanced technique, but to understand the idea behind exploiting a weakness in a controlled environment.</p><p>This helped me see that even basic vulnerabilities can create impact when an application does not properly protect sensitive actions.</p><p>For me, the main lesson was that security is not only about hiding pages. It is about validating permissions, protecting actions, and making sure users cannot perform operations they should not be allowed to perform.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>BANK-HACKED</em></blockquote><h3>What I Learned</h3><p>This was one of my first steps into offensive security.</p><p>To be honest, this module was very basic. I did not learn anything extremely advanced here, but that is completely fine because this is the beginning of the journey.</p><p>Sometimes we need to start with the basics of the basics.</p><p>This study helped me understand that offensive security is about controlled practice, curiosity, and learning how attackers think so we can build better defenses.</p><p>I also understood that web applications can expose information in different ways, and that hidden pages or poorly protected actions can create security problems.</p><h3>My Learning Journey</h3><p>I am a beginner ethical hacker, and this is part of my first cybersecurity learning path.</p><p>My goal is to keep studying, practicing, and documenting the process along the way.</p><p>I am also still learning English, and one of my goals is to become more fluent while studying cybersecurity during this next year.</p><p>If you are also learning cybersecurity or English, feel free to follow this journey with me.</p><h3>Final Thoughts</h3><p>This was a simple introduction, but every journey starts somewhere.</p><p>Even the basic topics matter because they help build the foundation for more advanced concepts later.</p><p>The more I study, the more I understand that cybersecurity is a long-term journey.</p><p>Let’s evolve together.</p><h3>tags</h3><p>Cybersecurity, Offensive Security, Ethical Hacking, Web Security, Cybersecurity Beginner, Learning Notes, Red Team, Blue Team, Information Security, Study Notes</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=36567bfa8c8d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Threat Modelling: My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/ai-threat-modelling-tryhackme-8961f3706ca1?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/8961f3706ca1</guid>
            <category><![CDATA[ai-threat-modelling]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Fri, 17 Apr 2026 11:21:01 GMT</pubDate>
            <atom:updated>2026-07-03T21:02:53.814Z</atom:updated>
            <content:encoded><![CDATA[<p>Warning: For educational purposes only.</p><p>When I started studying AI security, I realized that threat modelling for artificial intelligence systems is different from threat modelling traditional applications.</p><p>AI systems are not made only of code, servers, and databases. They can also include models, prompts, embeddings, training data, inference endpoints, vector databases, model registries, and automated decision-making workflows.</p><p>That means we need to think about new assets, new attack surfaces, and new ways attackers can abuse the system.</p><p>In this article, I want to share my beginner-friendly notes about AI threat modelling, AI-specific assets, STRIDE gaps, MITRE ATLAS, OWASP LLM risks, and practical threat modelling for AI assistants.</p><h3>About AI-Specific Assets and Attack Surfaces</h3><p>In a traditional system, we usually think about assets like servers, databases, APIs, users, and credentials.</p><p>In AI systems, the list becomes broader.</p><p>For example, in a RAG-based system, the application may retrieve relevant information at query time to help the model generate better answers. One important asset involved in this process is <strong>embedding vectors</strong>.</p><p>Embedding vectors are numerical representations of data. They help the system search and retrieve relevant context from a knowledge base or vector database.</p><p>Another important AI-specific asset is the <strong>model registry</strong>.</p><p>The model registry stores and manages model versions and artifacts. If an attacker compromises it and replaces the production model with a modified one, the impact can be serious.</p><p>This helped me understand that AI security is not only about protecting the final application. We also need to protect the data, the model lifecycle, the registry, and the systems that support inference.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Embedding Vectors”</em></blockquote><blockquote><em>“Model Registry / Artifacts”</em></blockquote><h3>About the Data Supply Chain and STRIDE’s Gaps</h3><p>AI systems depend heavily on data.</p><p>That means the data supply chain becomes a critical part of the security model.</p><p>One example is data poisoning. Instead of attacking the application directly, an attacker may inject crafted data points into the training pipeline over time. This can gradually influence the model’s behavior and shift its decision boundaries.</p><p>In this case, the malicious data is introduced during <strong>data collection</strong>.</p><p>This is different from many traditional application attacks because the impact may be delayed, subtle, and difficult to detect.</p><p>The STRIDE model is useful for threat modelling, but AI introduces some challenges. For example, the STRIDE category <strong>tampering</strong> may not fully capture the delayed and diffuse effects of training data poisoning.</p><p>For me, the key lesson is that AI threat modelling needs to consider time, data quality, model behavior, and indirect impact.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Data Collection”</em></blockquote><blockquote><em>“Tampering”</em></blockquote><h3>About Adapting STRIDE for AI Systems</h3><p>STRIDE is a common threat modelling framework, but AI systems require some adaptations.</p><p>In AI environments, <strong>Information Disclosure</strong> can appear in different ways. One important AI-specific manifestation is <strong>model extraction</strong>.</p><p>Model extraction happens when an attacker tries to steal or recreate a model by querying it and observing its outputs.</p><p>Another important risk is prompt-based abuse.</p><p>If an attacker crafts prompts that make an LLM bypass safety guidelines or content restrictions, this maps to <strong>Elevation of Privilege</strong> in a STRIDE-AI perspective.</p><p>This makes sense because the attacker is trying to make the system operate beyond its intended permissions or restrictions.</p><p>Another important AI risk is excessive agency.</p><p>In the OWASP LLM Top 10 2025, the risk of AI systems being granted too many permissions or too much autonomy is covered under <strong>LLM06: 2025 — Excessive Agency</strong>.</p><p>There is also a financial abuse scenario called <strong>Denial of Wallet</strong>.</p><p>This happens when an attacker does not necessarily take the service offline, but causes a huge increase in usage costs. For example, repeatedly sending expensive inference requests can drive cloud or API bills to extreme values.</p><p>This section helped me understand that availability in AI systems is not only about uptime. Cost exhaustion can also be a serious security concern.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Model Extraction”</em></blockquote><blockquote><em>“Elevation of Privilege”</em></blockquote><blockquote><em>“LLM06: 2025 — Excessive Agency”</em></blockquote><blockquote><em>“Denial of Wallet”</em></blockquote><h3>About MITRE ATLAS</h3><p>MITRE ATLAS is an important resource for understanding adversarial techniques against AI and machine learning systems.</p><p>ATLAS stands for <strong>Adversarial Threat Landscape for Artificial-Intelligence Systems</strong>.</p><p>It is similar in spirit to MITRE ATT&amp;CK, but focused on AI and ML threats.</p><p>One case study that caught my attention was <strong>Morris II</strong>.</p><p>This case described a self-replicating prompt injection worm that could spread between AI agents through RAG email systems.</p><p>This is a powerful example because it shows that prompt injection is not always isolated. In connected AI systems, malicious instructions can move between tools, agents, documents, emails, and retrieval systems.</p><p>Another important point from this section was the ATLAS technique ID for model extraction: <strong>AML.T0024</strong>.</p><p>The main lesson here is that AI threats are becoming structured enough to require their own knowledge bases, techniques, and defensive mappings.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Adversarial Threat Landscape for Artificial-Intelligence Systems”</em></blockquote><blockquote><em>“Morris II”</em></blockquote><blockquote><em>“AML.T0024”</em></blockquote><h3>About OWASP LLM Top 10 and Component Mapping</h3><p>The OWASP LLM Top 10 helps organize common risks in applications that use large language models.</p><p>One important part of AI threat modelling is mapping risks to components.</p><p>For example, the <strong>LLM Inference Endpoint</strong> is affected by several OWASP LLM Top 10 entries. In this study, the number was <strong>6</strong>.</p><p>This makes sense because the inference endpoint is where users interact with the model. It can be exposed to prompt injection, excessive resource usage, unsafe output handling, sensitive information disclosure, and other risks.</p><p>Another important example is <strong>Improper Output Handling</strong>.</p><p>If a chatbot renders LLM output directly in the browser without sanitization, the application may create risks such as script injection or unsafe content rendering.</p><p>This is a reminder that LLM output should not be blindly trusted.</p><p>Another key component is the <strong>training pipeline</strong>.</p><p>For risks related to data and model supply chain issues, such as LLM03, the training pipeline is one of the main components that needs hardening.</p><p>The main idea is simple: each component has its own risks. A secure AI system needs protection across the full architecture, not only at the chatbot interface.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“6”</em></blockquote><blockquote><em>“Improper Output Handling”</em></blockquote><blockquote><em>“Training Pipeline”</em></blockquote><h3>About the Practical Threat Modelling Challenge</h3><p>The practical part focused on threat modelling an AI assistant scenario.</p><p>This helped connect the theory with a more realistic system design.</p><p>Instead of only memorizing risks, the goal was to think about how an attacker could abuse the AI assistant, its tools, its data, its permissions, and its output.</p><p>For public notes, I prefer replacing the original platform acronym with a generic placeholder while keeping the structure understandable.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“ACRONYM_HERE{AI_THREAT_MODEL_COMPLETE}”</em></blockquote><h3>Why AI Threat Modelling Matters</h3><p>AI systems are becoming more integrated into business processes.</p><p>They can read documents, query databases, summarize emails, generate code, interact with APIs, trigger workflows, and support decision-making.</p><p>This creates value, but it also creates risk.</p><p>A threat model helps us ask important questions:</p><p>What are the AI-specific assets?</p><p>Where does user input enter the system?</p><p>What data is retrieved?</p><p>What tools can the model call?</p><p>What permissions does the system have?</p><p>Can model output trigger real actions?</p><p>Can attackers poison data?</p><p>Can attackers extract the model?</p><p>Can attackers abuse inference costs?</p><p>Can unsafe output reach users?</p><p>These questions help security teams understand the system before something goes wrong.</p><h3>What I Learned</h3><p>This study helped me understand that AI threat modelling requires a broader view of security.</p><p><strong>Embedding Vectors</strong> are important assets in RAG-based systems.</p><p><strong>Model Registry / Artifacts</strong> must be protected because compromised models can affect production behavior.</p><p><strong>Data Collection</strong> can be abused in data poisoning attacks.</p><p>Traditional <strong>Tampering</strong> may not fully explain the delayed effects of training data poisoning.</p><p><strong>Model Extraction</strong> is an AI-specific manifestation of Information Disclosure.</p><p>Prompt abuse can map to <strong>Elevation of Privilege</strong>.</p><p><strong>LLM06: 2025 — Excessive Agency</strong> covers risks related to too much permission or autonomy.</p><p><strong>Denial of Wallet</strong> is a cost-exhaustion attack against AI systems.</p><p><strong>ATLAS</strong> provides a knowledge base for AI and ML adversarial techniques.</p><p><strong>Improper Output Handling</strong> is a major risk when LLM output is rendered without proper sanitization.</p><p>The biggest takeaway for me is that AI systems need security at every layer: data, model, prompt, tools, APIs, outputs, permissions, and monitoring.</p><h3>Defensive Takeaways</h3><p>From a defensive perspective, this topic reinforced several important practices.</p><p>Protect model registries and artifacts.</p><p>Validate and monitor training data.</p><p>Harden the training pipeline.</p><p>Treat LLM output as untrusted.</p><p>Sanitize output before rendering it in browsers.</p><p>Apply least privilege to AI tools and agents.</p><p>Monitor inference endpoint usage.</p><p>Apply rate limits and cost controls.</p><p>Map AI risks using OWASP LLM Top 10.</p><p>Use MITRE ATLAS to understand adversarial AI techniques.</p><p>Review RAG systems for prompt injection and data poisoning risks.</p><p>Threat modelling helps teams find design weaknesses before attackers do.</p><h3>Final Thoughts</h3><p>Learning AI threat modelling helped me understand that AI security is not only about prompts.</p><p>It is about the full system.</p><p>The model is only one part of the architecture. Around it, there are data pipelines, embeddings, tools, registries, endpoints, permissions, logs, and user interfaces.</p><p>As a beginner, I am still learning, but this topic helped me see how important it is to think about AI security from the design phase.</p><p>The better we understand the system, the better we can protect it.</p><p>Thank you for reading this far.</p><p>Let’s evolve together.</p><h3>tags</h3><p>Cybersecurity, AI Security, Threat Modelling, LLM Security, OWASP, MITRE ATLAS, MLSecOps, Artificial Intelligence, Prompt Security, Blue Team, SOC, Defensive Security, Information Security, Cybersecurity Beginner, Learning Notes</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8961f3706ca1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Securing AI Systems: My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/securing-ai-systems-tryhackme-8aa618ec5d29?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/8aa618ec5d29</guid>
            <category><![CDATA[securing-ai-systems]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Fri, 17 Apr 2026 11:11:01 GMT</pubDate>
            <atom:updated>2026-07-03T21:00:38.601Z</atom:updated>
            <content:encoded><![CDATA[<p>Warning: For educational purposes only.</p><p>When I started studying artificial intelligence from a security perspective, I realized that AI systems are not just “a model answering questions.”</p><p>Behind an AI assistant, there may be prompts, tools, databases, APIs, retrieved context, permissions, logs, and automated actions.</p><p>That means AI security is not only about the model itself. It is about the entire system around it.</p><p>In this article, I want to share my beginner-friendly notes about AI system architecture, attack surfaces, system-level threats, security design patterns, and AI auditing.</p><h3>About the Anatomy of an AI System</h3><p>An AI system is usually made of several layers working together.</p><p>The model is an important part, but before the model receives anything, the system may need to prepare the prompt.</p><p>This is where <strong>prompt construction</strong> becomes important.</p><p>Prompt construction is responsible for combining the system prompt, the user input, and any retrieved context before sending everything to the model.</p><p>In simple words, it is the step that builds the final message the model will process.</p><p>Another important idea is the boundary between the LLM and external tools.</p><p>When an LLM output triggers something outside the model, such as a database query or an API call, it crosses a trust boundary. This matters because the model output is no longer just text. It can influence real actions.</p><p>In this study, that boundary was described as <strong>LLM-to-tools</strong>.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Prompt Construction”</em></blockquote><blockquote><em>“LLM-to-tools”</em></blockquote><h3>About the AI Attack Surface</h3><p>The attack surface of an AI system can be much larger than it first appears.</p><p>An attacker may not only target the model. They may target prompts, retrieved documents, plugins, APIs, databases, tools, permissions, logs, and output handling.</p><p>One important risk appears when LLM output is used by backend systems without proper validation.</p><p>For example, if model output is passed into a database query in an unsafe way, it may contribute to SQL injection risks. In the OWASP LLM Top 10 2025, this type of risk is covered under <strong>LLM05</strong>.</p><p>Another important resource for studying AI-related threats is <strong>MITRE ATLAS</strong>.</p><p>MITRE ATLAS is a knowledge base focused on adversary tactics and techniques against AI and machine learning systems.</p><p>For me, the big lesson here is that AI systems need security thinking from the beginning. We cannot treat model output as automatically safe just because it came from an AI assistant.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“LLM05”</em></blockquote><blockquote><em>“ATLAS”</em></blockquote><h3>About System-Level Threats</h3><p>System-level threats happen when the AI system interacts with real components, real data, or real actions.</p><p>One interesting case often discussed in AI risk conversations is the Air Canada chatbot incident. Although some people mention it as an LLM05 example, OWASP LLM Top 10 2025 classifies it under <strong>LLM09</strong>.</p><p>This helped me understand that AI risks are not always only technical injection problems. Sometimes the issue is misinformation, overreliance, or a system producing responses that create real-world consequences.</p><p>Another important concept is <strong>excessive agency</strong>.</p><p>Excessive agency happens when an AI system has too much ability to act, too many permissions, or too much autonomy.</p><p>The three dimensions are:</p><p><strong>Excessive Functionality</strong>, when the system can do more than it really needs.</p><p><strong>Excessive Permissions</strong>, when the system has more access than required.</p><p><strong>Excessive Autonomy</strong>, when the system can act without enough human approval or oversight.</p><p>Another risk appears when a user extracts internal details from an AI assistant, such as system prompts or internal API endpoints. In the OWASP LLM Top 10 2025, this falls under <strong>LLM07</strong>.</p><p>There is also the risk of resource abuse.</p><p>If an attacker sends thousands of maximum-length requests to an LLM API to generate a large bill or consume resources, this is covered under <strong>LLM10</strong>.</p><p>The key lesson here is that AI security also involves cost protection, access control, system design, and operational limits.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“LLM09”</em></blockquote><blockquote><em>“Excessive Functionality, Excessive Permissions, Excessive Autonomy”</em></blockquote><blockquote><em>“LLM07”</em></blockquote><blockquote><em>“LLM10”</em></blockquote><h3>About Security Design Patterns</h3><p>Security design patterns help us build safer AI systems.</p><p>One of the most important principles is <strong>Least Privilege</strong>.</p><p>Least privilege means that every component should only have the minimum permissions required to perform its function.</p><p>This is extremely important in AI systems because an assistant may interact with tools, APIs, repositories, databases, or internal systems.</p><p>If the AI component has unnecessary permissions, a mistake or attack can cause much more damage.</p><p>Another important practice is <strong>MLSecOps</strong>.</p><p>MLSecOps integrates security into the machine learning lifecycle. This includes monitoring, observability, incident response, and continuous improvement.</p><p>In simple words, MLSecOps means we should not think about AI security only after deployment. Security should be part of how AI systems are designed, built, monitored, and maintained.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Least Privilege”</em></blockquote><blockquote><em>“MLSecOps”</em></blockquote><h3>About Auditing an AI Assistant</h3><p>The auditing part helped me think like a security reviewer.</p><p>When auditing an AI assistant, we need to understand what actions it can perform, what permissions it has, what data it can access, and what security controls are missing.</p><p>One risky behavior identified in the case study was that the assistant could automatically <strong>merge pull requests</strong> without requiring human approval.</p><p>That is a big concern because merging code is a sensitive action. If an AI system can do this automatically, a mistake or malicious manipulation could affect the codebase.</p><p>Another important finding was the database role.</p><p>The assistant reported operating under <strong>db_admin</strong>, which suggests excessive database privileges.</p><p>From a security perspective, this violates the principle of least privilege. An AI assistant should not have administrative database access unless there is a very strong reason and strict controls around it.</p><p>The final issue was related to logging.</p><p>The assistant logged all conversations without applying <strong>PII Filtering</strong>.</p><p>This is risky because conversations may contain personal information, secrets, internal data, credentials, or sensitive business details.</p><p>A secure AI system should think carefully about what is logged, where logs are stored, who can access them, and how sensitive data is filtered or protected.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Merge Pull Requests”</em></blockquote><blockquote><em>“db_admin”</em></blockquote><blockquote><em>“PII Filtering”</em></blockquote><h3>Why This Matters for Security</h3><p>AI systems are becoming more connected to real workflows.</p><p>They can summarize documents, query databases, trigger tools, open tickets, write code, review pull requests, and interact with internal systems.</p><p>This makes AI useful, but it also creates risk.</p><p>If an AI system has too much permission, trusts model output blindly, logs sensitive data, or triggers actions without review, it can become a serious security problem.</p><p>Some important security questions are:</p><p>What can the AI system access?</p><p>What can it change?</p><p>What tools can it call?</p><p>Can it trigger real actions?</p><p>Does it require human approval?</p><p>Are logs filtered for sensitive data?</p><p>Are permissions limited?</p><p>Is output validated before being used?</p><p>For me, this topic made one thing very clear: AI security is system security.</p><h3>What I Learned</h3><p>This study helped me understand that securing AI systems requires looking beyond the model.</p><p><strong>Prompt Construction</strong> combines the system prompt, user input, and retrieved context before sending it to the model.</p><p><strong>LLM-to-tools</strong> is an important boundary because model output can trigger real actions.</p><p><strong>LLM05</strong> covers risks where LLM output can lead to backend issues such as SQL injection.</p><p><strong>ATLAS</strong> is the MITRE knowledge base for adversary tactics and techniques against AI and ML systems.</p><p><strong>LLM09</strong> can apply to cases involving misinformation or problematic model behavior with real-world impact.</p><p><strong>Excessive Functionality, Excessive Permissions, and Excessive Autonomy</strong> are the three dimensions of excessive agency.</p><p><strong>LLM07</strong> covers risks related to exposing sensitive system information.</p><p><strong>LLM10</strong> covers resource abuse and cost-related attacks.</p><p><strong>Least Privilege</strong> helps reduce the impact of compromise or misuse.</p><p><strong>MLSecOps</strong> brings security into the machine learning lifecycle.</p><p>The biggest takeaway for me is that AI systems need strong boundaries, limited permissions, monitoring, and human oversight.</p><h3>Defensive Takeaways</h3><p>From a defensive perspective, this topic reinforces several important practices.</p><p>Apply least privilege to AI components.</p><p>Validate model output before using it in backend systems.</p><p>Avoid giving AI assistants unnecessary database permissions.</p><p>Require human approval for sensitive actions.</p><p>Filter PII and sensitive data from logs.</p><p>Monitor AI tool usage.</p><p>Limit request size and rate to reduce abuse.</p><p>Review system prompts and tool access.</p><p>Treat LLM-to-tools boundaries as security-sensitive.</p><p>Build security into the AI lifecycle through MLSecOps.</p><p>A secure AI system is not just a smart model. It is a controlled, monitored, and well-designed environment.</p><h3>Final Thoughts</h3><p>Learning about AI system security helped me see artificial intelligence with a more realistic view.</p><p>AI assistants can be powerful, but they also introduce new risks when connected to tools, data, APIs, and automated workflows.</p><p>As a beginner, I am still learning, but this topic helped me understand that securing AI means securing the full architecture.</p><p>The better we understand the system, the better we can protect it.</p><p>Thank you for reading this far.</p><p>Let’s evolve together.</p><h3>tags</h3><p>Cybersecurity, AI Security, LLM Security, OWASP, MITRE ATLAS, MLSecOps, Prompt Security, Artificial Intelligence, Blue Team, SOC, Defensive Security, Information Security, Cybersecurity Beginner, Learning Note</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8aa618ec5d29" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Prompt Engineering: My Beginner-Friendly Notes]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/prompt-engineering-tryhackme-ea73a00b499c?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/ea73a00b499c</guid>
            <category><![CDATA[prompt-engineering]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Fri, 17 Apr 2026 11:06:01 GMT</pubDate>
            <atom:updated>2026-07-03T20:56:57.740Z</atom:updated>
            <content:encoded><![CDATA[<h3>Warning: For educational purposes only.</h3><p>When I started learning about artificial intelligence and cybersecurity, I realized that knowing how to ask good questions is becoming a very important skill.</p><p>Prompt engineering is not just about writing random instructions to an AI model. It is about learning how to communicate clearly, provide context, define rules, and guide the model toward a useful answer.</p><p>In this article, I want to share my beginner-friendly notes about LLM fundamentals, prompt structure, system and user prompts, and some advanced prompting techniques.</p><h3>About LLM Fundamentals</h3><p>Before learning prompt engineering, it is important to understand some basic concepts about Large Language Models, also known as LLMs.</p><p>An LLM processes text by breaking it into smaller units called <strong>tokens</strong>. These tokens can be words, parts of words, symbols, or characters, depending on how the model reads the input.</p><p>Another important concept is <strong>temperature</strong>.</p><p>Temperature controls how creative or predictable the model’s answer can be. When the temperature is set closer to <strong>0.0</strong>, the model tends to behave in a more deterministic way, meaning it gives more consistent and predictable answers.</p><p>There is also a parameter called <strong>top-p</strong>.</p><p>Top-p controls which tokens the model considers by limiting the selection to a cumulative probability mass. In simple words, it helps control how broad or narrow the model’s choices can be when generating text.</p><p>Another key concept is the <strong>context window</strong>.</p><p>The context window is the maximum working memory of the model, measured in tokens. It defines how much information the model can consider at once during a conversation or task.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Tokens”</em></blockquote><blockquote><em>“Temperature”</em></blockquote><blockquote><em>“Top-p”</em></blockquote><blockquote><em>“Context window”</em></blockquote><h3>About the Anatomy of a Prompt</h3><p>A good prompt usually has more than just one simple instruction.</p><p>It can include the task, the context, the expected format, and the rules the model should follow.</p><p>One important part is the <strong>instruction</strong>. This is the core command or action we want the AI to perform.</p><p>Another important part is the <strong>context</strong>. Context gives the model background information so it can better understand the situation.</p><p>There are also <strong>constraints</strong>. Constraints define rules or limits for the answer, such as tone, length, style, or topics that should be avoided.</p><p>Finally, the <strong>output format</strong> tells the model how the answer should be structured. For example, the answer could be written as bullet points, a table, a JSON object, a short paragraph, or a step-by-step explanation.</p><p>The way I understood it is simple:</p><p>A good prompt is like giving clear instructions to someone who wants to help, but needs to know exactly what you expect.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Output Format”</em></blockquote><blockquote><em>“Constraints”</em></blockquote><blockquote><em>“Context”</em></blockquote><blockquote><em>“Instruction”</em></blockquote><h3>About System and User Prompts</h3><p>Another important topic is the difference between system prompts and user prompts.</p><p>A <strong>system prompt</strong> is usually developer-defined, persistent, and remains constant across sessions. It helps define how the model should behave inside a specific application.</p><p>A <strong>user prompt</strong>, on the other hand, is the message or instruction written by the person interacting with the AI.</p><p>One concept that helped me understand this better is <strong>instruction hierarchy</strong>.</p><p>Instruction hierarchy defines the intended order of priority between system and user instructions in an LLM application.</p><p>In simple words, not all instructions have the same level of authority. Some instructions are designed to guide the model at a deeper level, while others are temporary requests from the user.</p><p>This matters because AI applications need structure, safety, and predictable behavior.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“System prompt”</em></blockquote><blockquote><em>“Instruction hierarchy”</em></blockquote><h3>About Advanced Prompting Techniques</h3><p>Prompt engineering also includes different techniques that can improve how we interact with AI models.</p><p>One well-known technique is <strong>Chain-of-Thought</strong>. This technique became popular after being introduced by Google researchers in 2022, and it focuses on asking models to break tasks into intermediate reasoning steps.</p><p>Another technique is <strong>zero-shot prompting</strong>.</p><p>Zero-shot means asking the model to complete a task without giving examples. In this case, the model relies on its pre-trained knowledge and the instruction provided in the prompt.</p><p>There are also <strong>prompt templates</strong>.</p><p>Prompt templates are standardized prompt structures that can be saved and reused for recurring tasks. This is useful when we want consistency, especially when performing similar tasks many times.</p><p>Another simple phrase often associated with zero-shot reasoning is:</p><p>“Let’s think step by step.”</p><p>This phrase can encourage the model to approach a task more carefully and organize the answer in a more structured way.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“Chain-of-Thought”</em></blockquote><blockquote><em>“Zero-shot”</em></blockquote><blockquote><em>“Prompt templates”</em></blockquote><blockquote><em>“Let’s think step by step”</em></blockquote><h3>About the Challenge</h3><p>The final part was a simple challenge to validate the concepts learned during the study.</p><p>Instead of focusing only on memorizing terms, this part helped reinforce the importance of understanding how prompts are structured and how different settings can affect the model’s behavior.</p><p>For public notes, I prefer replacing the platform acronym with a generic placeholder. This keeps the format understandable without directly exposing the original platform tag.</p><p><strong>My learning notes from this section:</strong></p><blockquote><em>“ACRONYM_HERE{Pr0mpt_3ng1neer}”</em></blockquote><h3>Why Prompt Engineering Matters</h3><p>Prompt engineering is becoming useful in many areas.</p><p>For cybersecurity, it can help with writing reports, summarizing logs, creating detection ideas, explaining technical concepts, building study notes, and improving communication.</p><p>But it also requires responsibility.</p><p>A good prompt should be clear, ethical, and aligned with the task. The goal is not only to get an answer, but to get an answer that is useful, structured, and safe.</p><p>For beginners, prompt engineering is a great skill because it teaches us how to think clearly before asking for help.</p><h3>What I Learned</h3><p>This study helped me understand that prompt engineering is about clarity.</p><p><strong>Tokens</strong> are the small units that an LLM uses to process text.</p><p><strong>Temperature</strong> controls how deterministic or creative the model can be.</p><p><strong>Top-p</strong> limits token selection based on probability.</p><p>The <strong>context window</strong> defines how much information the model can work with at once.</p><p>A good prompt can include <strong>instruction</strong>, <strong>context</strong>, <strong>constraints</strong>, and <strong>output format</strong>.</p><p>A <strong>system prompt</strong> is persistent and usually defined by the developer.</p><p><strong>Instruction hierarchy</strong> defines the priority order between different types of instructions.</p><p>Advanced techniques like <strong>Chain-of-Thought</strong>, <strong>zero-shot prompting</strong>, and <strong>prompt templates</strong> can help improve results when used correctly.</p><p>The biggest lesson for me is that better prompts usually create better answers.</p><h3>Final Thoughts</h3><p>Learning prompt engineering helped me see AI tools with more structure.</p><p>Instead of just typing a question and hoping for a good answer, we can guide the model by being clear about the task, the context, the rules, and the expected format.</p><p>As a beginner, I am still learning both cybersecurity and English, but this topic showed me how important communication is in technology.</p><p>The better we communicate, the better we learn.</p><p>Thank you for reading this far.</p><p>Let’s evolve together.</p><h3>tags</h3><p>Cybersecurity, Prompt Engineering, Artificial Intelligence, LLM, AI Security, Learning Notes, Cybersecurity Beginner, Technology, Digital Skills, Information Security, Study Notes</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ea73a00b499c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Tryhackme Web Application Basics]]></title>
            <link>https://medium.com/@andreyna.m.carvalho/tryhackme-web-application-basics-2b4c9da1ee96?source=rss-a477eba5584e------2</link>
            <guid isPermaLink="false">https://medium.com/p/2b4c9da1ee96</guid>
            <category><![CDATA[basics]]></category>
            <category><![CDATA[web-applications]]></category>
            <category><![CDATA[web-application-basics]]></category>
            <dc:creator><![CDATA[Andreyna Marques Carvalho]]></dc:creator>
            <pubDate>Sun, 16 Nov 2025 11:43:07 GMT</pubDate>
            <atom:updated>2026-07-03T20:53:39.578Z</atom:updated>
            <content:encoded><![CDATA[<h3>Web Application Basics: My Beginner-Friendly Notes</h3><p>When I started studying web application security, I realized that before learning about vulnerabilities, attacks, and defenses, we first need to understand how web applications actually work.</p><p>A web application is something we interact with through a browser, but behind the screen there are many components working together: the browser, the web server, the database, HTTP messages, headers, cookies, status codes, and security controls.</p><p>This article is a simple summary of my beginner-friendly notes about web application basics, HTTP communication, URLs, request methods, response codes, cookies, and security headers.</p><h3>About Web Applications</h3><p>A web application is an application that runs through a web browser.</p><p>When we access a website, fill out a form, log in to a system, or use an online platform, we are usually interacting with a web application.</p><p>There are two main sides involved: the front end and the back end.</p><p>The <strong>front end</strong> is what we see and interact with in the browser. It usually includes HTML, CSS, and JavaScript.</p><p>The <strong>back end</strong> is what works behind the scenes. It can include the web server, database, application logic, and security layers.</p><p>One important component is the <strong>web server</strong>, which is responsible for hosting and delivering content for web applications.</p><p>The tool we use to access and interact with web applications is the <strong>web browser</strong>.</p><p>Another important security component is the <strong>Web Application Firewall</strong>, also known as WAF. It acts as a protective layer by filtering incoming traffic and helping block malicious requests.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Web server</blockquote><blockquote>Web browser</blockquote><blockquote>Web application firewall</blockquote><h3>About URLs</h3><p>A URL is what we use to access a resource on the web.</p><p>It tells the browser where to go, which protocol to use, what domain to contact, and sometimes what extra information should be sent to the server.</p><p>One important protocol is <strong>HTTPS</strong>.</p><p>HTTPS provides encrypted communication between a web browser and a web server. This helps protect data during transmission, especially when users send sensitive information like passwords, tokens, or personal details.</p><p>Another interesting concept is <strong>typosquatting</strong>.</p><p>Typosquatting happens when someone registers misspelled versions of popular domain names to trick users. This can be used for phishing, fraud, malware distribution, or fake login pages.</p><p>URLs can also include a <strong>query string</strong>.</p><p>A query string is used to pass additional information to the web server, such as search terms, filters, tracking values, or form inputs.</p><p><strong>My learning notes from this section:</strong></p><blockquote>HTTPS</blockquote><blockquote>Typosquatting</blockquote><blockquote>Query String</blockquote><h3>About HTTP Messages</h3><p>HTTP is the protocol used for communication between a client and a web server.</p><p>In simple words, the browser sends an HTTP request, and the server returns an HTTP response.</p><p>An <strong>HTTP response</strong> is the message returned by the web server after processing the client’s request.</p><p>HTTP messages usually contain headers, and after the headers there is an <strong>empty line</strong>. This empty line separates the headers from the message body.</p><p>This structure is important because it helps the client and server understand where metadata ends and where the actual content begins.</p><p><strong>My learning notes from this section:</strong></p><blockquote>HTTP response</blockquote><blockquote>Empty Line</blockquote><h3>About HTTP Request Lines and Methods</h3><p>An HTTP request tells the server what the client wants to do.</p><p>For example, the client may want to retrieve data, send data, update something, or delete something.</p><p>One important HTTP version is <strong>HTTP/1.1</strong>.</p><p>HTTP/1.1 became widely adopted and remains one of the most common versions used in web communication. It introduced important features such as persistent connections and chunked transfer encoding.</p><p>HTTP also has different request methods.</p><p>The <strong>OPTIONS</strong> method is used to describe the communication options available for a target resource. In simple words, it can help identify which HTTP methods are supported by the web server.</p><p>Another important part of an HTTP request is the <strong>URL Path</strong>.</p><p>The URL path specifies the specific resource or endpoint being requested on the web server. It usually appears after the domain name.</p><p>For example, in a simple URL, the path may point to a page, an API endpoint, or a specific resource.</p><p><strong>My learning notes from this section:</strong></p><blockquote>HTTP/1.1</blockquote><blockquote>OPTIONS</blockquote><blockquote>URL Path</blockquote><h3>About HTTP Request Headers and Body</h3><p>HTTP request headers contain additional information that helps the server understand how to process the request.</p><p>One important request header is <strong>Host</strong>.</p><p>The Host header specifies the domain name of the web server to which the request is being sent.</p><p>Another important concept is the content type used in form submissions.</p><p>The default content type for form submissions where data is encoded as key=value pairs is <strong>application/x-www-form-urlencoded</strong>.</p><p>This format is very common when forms send simple data to a server.</p><p>The <strong>request headers</strong> are important because they can include information such as host, user agent, accepted content types, content length, cookies, authorization data, and content type.</p><p>From a security perspective, headers matter because they can affect authentication, session handling, content processing, and application behavior.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Host</blockquote><blockquote>application/x-www-form-urlencoded</blockquote><blockquote>Request Headers</blockquote><h3>About HTTP Response Status Codes</h3><p>When a server responds to a request, it includes information about the result.</p><p>The <strong>Status Line</strong> is the part of an HTTP response that provides the HTTP version, status code, and a short explanation of the response outcome.</p><p>Status codes are grouped into categories.</p><p>For example, <strong>Server Error Responses</strong> indicate that the web server encountered an internal issue or was unable to fulfill the client’s request.</p><p>One of the most famous status codes is <strong>404</strong>.</p><p>A 404 status code means that the requested resource could not be found on the web server.</p><p>This is the kind of response we usually see when a page does not exist or the URL path is incorrect.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Status Line</blockquote><blockquote>Server Error Responses</blockquote><blockquote>404</blockquote><h3>About HTTP Response Headers and Body</h3><p>HTTP response headers can reveal important information about the server and how the browser should handle the response.</p><p>One header that deserves attention is the <strong>Server</strong> header.</p><p>The Server header can reveal information about the web server software and version. From a security perspective, exposing this information may help attackers identify known vulnerabilities related to that specific version.</p><p>Cookies are also commonly configured through HTTP response headers.</p><p>The <strong>Secure</strong> flag should be added to cookies to ensure they are only transmitted over HTTPS. This helps protect cookies from being exposed over unencrypted connections.</p><p>The <strong>HttpOnly</strong> flag should be added to cookies to prevent them from being accessed through JavaScript. This helps reduce the risk of cookie theft in some Cross-Site Scripting scenarios.</p><p>These small configurations can make a big difference in web application security.</p><p><strong>My learning notes from this section:</strong></p><blockquote>Server</blockquote><blockquote>Secure</blockquote><blockquote>HttpOnly</blockquote><h3>About Security Headers</h3><p>Security headers help browsers enforce safer behavior when loading and interacting with a website.</p><p>One important security header is <strong>Content Security Policy</strong>, also known as CSP.</p><p>Inside a CSP configuration, the <strong>script-src</strong> directive can define where scripts are allowed to be loaded from. This helps reduce the risk of loading malicious scripts from untrusted sources.</p><p>Another important header is <strong>Strict-Transport-Security</strong>, also known as HSTS.</p><p>When configuring HSTS, the <strong>includeSubDomains</strong> directive can be used to apply the HTTPS-only policy to both the main domain and its subdomains.</p><p>There is also the <strong>nosniff</strong> directive.</p><p>The nosniff directive helps prevent browsers from interpreting files as a different MIME type than the one specified by the server. This can help reduce content type sniffing attacks.</p><p>Security headers are simple in concept, but they are very powerful when configured correctly.</p><p><strong>My learning notes from this section:</strong></p><blockquote>script-src</blockquote><blockquote>includeSubDomains</blockquote><blockquote>nosniff</blockquote><h3>About Making HTTP Requests</h3><p>The practical part helped me understand how different HTTP methods work in a simple API context.</p><p>A <strong>GET</strong> request is commonly used to retrieve information from a server.</p><p>A <strong>POST</strong> request is commonly used to send or update data.</p><p>A <strong>DELETE</strong> request is commonly used to remove a resource.</p><p>In an API, these methods help define the action being performed against a specific endpoint.</p><p>This was useful because it connected the theory of HTTP methods with practical behavior.</p><p>Instead of only memorizing what GET, POST, and DELETE mean, I could see how each method changes the interaction with the application.</p><p><strong>My learning notes from this section:</strong></p><blockquote>GET request result: ACRONYM_HERE{YOU_HAVE_JUST_FOUND_THE_USER_LIST}</blockquote><blockquote>POST request result: ACRONYM_HERE{YOU_HAVE_MODIFIED_THE_USER_DATA}</blockquote><blockquote>DELETE request result: ACRONYM_HERE{YOU_HAVE_JUST_DELETED_A_USER}</blockquote><h3>Why This Matters for Security</h3><p>Web application basics are extremely important for cybersecurity.</p><p>Many vulnerabilities happen because of how applications handle requests, responses, headers, cookies, sessions, input, and authentication.</p><p>Understanding HTTP helps when studying topics like:</p><p>Cross-Site Scripting</p><p>SQL Injection</p><p>Authentication flaws</p><p>Session hijacking</p><p>Broken access control</p><p>Security misconfiguration</p><p>API security</p><p>Web server hardening</p><p>Security headers</p><p>Cookie protection</p><p>Before learning how web attacks work, we need to understand normal web communication.</p><p>That is why learning the basics of web applications is such an important step.</p><h3>What I Learned</h3><p>This study helped me understand that web applications are built from many different components.</p><p>A <strong>web server</strong> hosts and delivers web content.</p><p>A <strong>web browser</strong> is used to access and interact with web applications.</p><p>A <strong>web application firewall</strong> helps filter malicious traffic.</p><p><strong>HTTPS</strong> provides encrypted communication.</p><p><strong>Typosquatting</strong> abuses misspelled domain names.</p><p>A <strong>query string</strong> passes additional information to the server.</p><p>An <strong>HTTP response</strong> is returned by the web server after processing a request.</p><p>An <strong>empty line</strong> separates headers from the message body.</p><p><strong>HTTP/1.1</strong> is a widely adopted HTTP version.</p><p><strong>OPTIONS</strong> describes supported communication methods.</p><p>The <strong>URL Path</strong> identifies the requested resource.</p><p>The <strong>Host</strong> header specifies the target domain.</p><p><strong>application/x-www-form-urlencoded</strong> is a common form submission content type.</p><p><strong>Request Headers</strong> guide how the server should process the request.</p><p>The <strong>Status Line</strong> shows the HTTP version, status code, and response explanation.</p><p><strong>Server Error Responses</strong> indicate internal server problems.</p><p><strong>404</strong> means the requested resource was not found.</p><p>The <strong>Server</strong> header may reveal server software and version.</p><p>The <strong>Secure</strong> cookie flag helps protect cookies over HTTPS.</p><p>The <strong>HttpOnly</strong> cookie flag helps protect cookies from JavaScript access.</p><p><strong>script-src</strong> controls where scripts can be loaded from in CSP.</p><p><strong>includeSubDomains</strong> applies HSTS to subdomains.</p><p><strong>nosniff</strong> helps prevent MIME type sniffing.</p><p>The biggest takeaway for me is that web security starts with understanding how the web works.</p><h3>Defensive Takeaways</h3><p>From a defensive perspective, this topic reinforces several important practices.</p><p>Use HTTPS everywhere.</p><p>Configure cookies with Secure and HttpOnly when appropriate.</p><p>Avoid exposing unnecessary server version information.</p><p>Use security headers such as CSP, HSTS, and nosniff.</p><p>Validate and monitor HTTP methods.</p><p>Review API endpoints carefully.</p><p>Understand status codes during troubleshooting and security analysis.</p><p>Use a WAF as an additional protective layer, but not as the only defense.</p><p>Keep web servers and frameworks updated.</p><p>Web application security is not only about finding bugs. It is also about building and configuring applications safely from the beginning.</p><h3>Final Thoughts</h3><p>Learning web application basics helped me understand the foundation behind many cybersecurity topics.</p><p>HTTP requests, responses, headers, methods, status codes, cookies, and security headers may look simple at first, but they are everywhere in web security.</p><p>As a beginner, I am still building my foundation, but this topic helped me connect web development concepts with cybersecurity thinking.</p><p>The better we understand how web applications work, the better we can understand how to protect them.</p><p>Let’s evolve together.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2b4c9da1ee96" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>