<?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 Victoria Arinola on Medium]]></title>
        <description><![CDATA[Stories by Victoria Arinola on Medium]]></description>
        <link>https://medium.com/@arinolatemilola?source=rss-30ccf3daeca5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*lcYMijRckZGyd1xOJnOiQg@2x.jpeg</url>
            <title>Stories by Victoria Arinola on Medium</title>
            <link>https://medium.com/@arinolatemilola?source=rss-30ccf3daeca5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 21:10:37 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@arinolatemilola/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[Your App Has Holes. Here’s How Strix, a Free AI Hacker, Will Find Them]]></title>
            <link>https://medium.com/@arinolatemilola/your-app-has-holes-heres-how-strix-a-free-ai-hacker-will-find-them-f475a2bf896d?source=rss-30ccf3daeca5------2</link>
            <guid isPermaLink="false">https://medium.com/p/f475a2bf896d</guid>
            <dc:creator><![CDATA[Victoria Arinola]]></dc:creator>
            <pubDate>Thu, 19 Mar 2026 22:16:35 GMT</pubDate>
            <atom:updated>2026-03-19T22:16:35.637Z</atom:updated>
            <content:encoded><![CDATA[<h3>What is Strix?</h3><p>Strix is an open-source, agentic security platform that deploys autonomous AI agents to act like real-world hackers. Unlike static analysers that flag patterns, Strix dynamically runs your code, probes your endpoints, and validates every vulnerability through actual proof-of-concept (PoC) exploits, giving you confirmed, high-signal findings instead of noise.</p><blockquote><strong>Why Strix for Security Professionals?</strong></blockquote><blockquote>Traditional pentests take weeks and cost tens of thousands of dollars. Static analysers flood you with false positives. Strix delivers confirmed PoC-validated vulnerabilities in hours, designed to integrate into your CI/CD pipeline and existing security workflow.</blockquote><h3>Key Capabilities</h3><p>• Autonomous multi-agent architecture: Specialised agents collaborate, share intel, and adapt attack strategies in real time</p><p>• Dynamic execution in a Docker sandbox: Runtime vulnerabilities missed by SAST tools are caught through actual code execution</p><p>• PoC validation: Every finding is proven through a real exploit attempt, eliminating false positives</p><p>• Multi-target support: local codebases, GitHub repositories, live web applications, or all three combined</p><p>• CI/CD integration: block vulnerabilities in pull requests before they reach production</p><p>• LLM-agnostic: OpenAI, Anthropic Claude, Google Vertex AI, AWS Bedrock, Azure, or local models</p><h3>Vulnerability Coverage</h3><p>• Access control flaws and IDOR (Insecure Direct Object Reference)</p><p>• SQL injection, command injection, template injection</p><p>• XSS, CSRF, and DOM-based vulnerabilities</p><p>• SSRF and server-side vulnerabilities</p><p>• Authentication and session management weaknesses</p><p>• Business logic flaws and race conditions</p><p>• Misconfigured attack surfaces and exposed debug endpoints</p><h3>Understanding Credits &amp; Token Costs</h3><p>This is one of the most important things to understand before running Strix. The Strix software itself is free and open-source (Apache 2.0 licence). However, Strix is powered entirely by Large Language Models (LLMs), and those LLMs charge for every token of input and output they process. Every time Strix reasons, plans an attack, interprets a tool output, or generates a PoC, it is consuming LLM API tokens. Those tokens cost real money.</p><blockquote><strong>⚠️ Strix is free. The AI brain powering it is not.</strong></blockquote><blockquote>Think of Strix as the engine and the LLM as the fuel. You supply the fuel via your API key and pay the LLM provider directly per token consumed. A standard scan on a medium-complexity application typically consumes between 40,000 and 100,000 tokens.</blockquote><h3>How Token Consumption Works</h3><p>Each Strix agent operates in a continuous reasoning loop. For every action it takes, mapping an endpoint, crafting a payload, interpreting a server response, it sends a prompt to the LLM and receives a response. Both the input prompt and the output response consume tokens. Complex targets with many endpoints, large codebases, or deep scan modes will naturally burn through more tokens.</p><p>Here is what drives token consumption during a scan:</p><p>• Attack planning: the orchestrator agent reads your codebase or probes endpoints and builds a testing strategy</p><p>• Tool output interpretation: every response from tools like Nmap, Nuclei, or the HTTP proxy is sent to the LLM for analysis</p><p>• Payload generation: crafting and iterating SQL injection strings, XSS vectors, IDOR permutations, etc.</p><p>• Proof-of-concept construction: the LLM writes and validates working exploit code for confirmed vulnerabilities</p><p>• Report generation: the final structured report is compiled by the LLM from all agent findings</p><h3>Which LLM Should You Use?</h3><p>Strix recommends using the highest-capability models for best vulnerability detection rates. Weaker or cheaper models will miss more vulnerabilities and produce lower-quality PoCs.</p><blockquote><strong>Best for accuracy: strix/gpt-5, strix/claude-opus-4–6, strix/gpt-5.2</strong></blockquote><blockquote>These are the recommended models via Strix Router. They deliver the highest detection rates and the clearest remediation guidance. For CI/CD quick scans where speed matters more than depth, strix/gpt-4o is a cost-effective choice.</blockquote><blockquote><strong>Cost-saving tip: Use scan modes strategically</strong></blockquote><blockquote>Run quick scans: (scan-mode quick) in CI/CD pipelines to catch high-severity issues cheaply on every PR.</blockquote><blockquote>Reserve deep scans: ( scan-mode deep) for pre-release security reviews or when specifically investigating a suspected vulnerability class.</blockquote><blockquote><strong>Local models: Free but lower accuracy</strong></blockquote><blockquote>You can run Strix against a local model via Ollama or LM Studio (set LLM_API_BASE). This eliminates API costs entirely but significantly reduces detection rates. Not recommended for production security assessments.</blockquote><h3>Managing Your API Spend</h3><p>• Set a hard spending cap on your LLM provider dashboard before running any scan</p><p>• Use strix/gpt-4o or similar mid-tier models for exploration; upgrade to gpt-5 for final deep assessments</p><p>• Run scan-mode quick first to estimate coverage, then deepen selectively with instructions to focus on specific attack surfaces</p><p>• For teams, Strix Cloud (app.strix.ai) offers a managed credit system with per-scan billing transparency</p><p>• Monitor token usage live in the Strix TUI, the ‘Tokens used’ counter updates in real time during a scan</p><h3>Prerequisites</h3><p>Before installing Strix on WSL (Ubuntu on Windows), ensure the following are in place:</p><h3>1. WSL with Ubuntu</h3><p>Open your Ubuntu terminal from the Windows Start menu or by running wsl in PowerShell/CMD. You should see a bash prompt like the one below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*x66pMCsmJkMY4ZKToD_IPg.png" /><figcaption><em>Ubuntu WSL2 terminal open and ready</em></figcaption></figure><h3>2. Python 3.12 or Newer</h3><p>Strix requires Python 3.12+. Ubuntu 22.04 ships with Python 3.10 by default. Check and upgrade if needed:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/940/1*vL1_815IhIqzVVUy9a1jvw.png" /></figure><h3>3. Docker Desktop (Running)</h3><p>Strix executes its security sandbox inside Docker containers. Docker Desktop must be installed on Windows and configured to integrate with WSL 2.</p><p>• Download Docker Desktop from docker.com/products/docker-desktop</p><p>• During install, enable the WSL 2 backend</p><p>• In Docker Desktop → Settings → Resources → WSL Integration, enable your Ubuntu distro</p><ul><li>Verify Docker is accessible from inside WSL:</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*ZUt_UD5xoZVSn4E6qs7kWw.png" /><figcaption><em>Docker accessible from WSL2, version confirmed</em></figcaption></figure><h3>4. An LLM API Key (Credits Required)</h3><p>Because Strix is powered by an LLM, you must have an active API key with sufficient credit balance before running a scan. As covered in the Credits &amp; Token Costs section above, scans consume tokens, ensure your account has credit loaded.</p><blockquote><strong>Option A — Strix Router (Recommended)</strong></blockquote><blockquote>Sign up at models.strix.ai. Load credits to your Strix account. You get access to top models (GPT-5, Claude Opus, GPT-5.2) at discounted rates, optimised for security testing. Credits are consumed per token as agents work.</blockquote><blockquote><strong>Option B — Bring Your Own Key (BYOK)</strong></blockquote><blockquote>Use an API key from OpenAI, Anthropic, Google, AWS, or Azure. Charges go directly to your provider account at their standard token rates. Ensure your account has a spending limit set to avoid surprises.</blockquote><h3>Installation &amp; Setup</h3><h3>Step 1 — Install Strix</h3><p>Open your Ubuntu WSL terminal and run the official install script:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/939/1*cdS9XRtbQksRXljxWHbMDA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*r8pyye7CitXw7NT_X_kQWw.png" /></figure><p>Alternatively, install via pipx if you manage Python tool environments:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/931/1*zvpy69e1aCMDpiBvS4Agvw.png" /></figure><h3>Verify the Installation</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/933/1*nTopBVyViCBDWbe-U2iflQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*8NCFgLL_8c44RJZ1HnwBfA.png" /><figcaption><em>Strix version confirmed after successful installation</em></figcaption></figure><h3>Step 2 — Configure Your LLM Provider</h3><p>Set your LLM model and API key as environment variables. Add them to ~/.bashrc to persist across WSL sessions:</p><h3>Strix Router (Recommended)</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/931/1*5jgMqXOR3cxuHu_YjYlE1Q.png" /></figure><h3>OpenAI / Anthropic / Others</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/933/1*47lgoapRKTOYN-RYuoFBQA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*vA39r4t7PTE8r6MCsVEWuQ.png" /><figcaption><em>LLM provider configured and verified in terminal</em></figcaption></figure><blockquote><strong>Persist config across sessions</strong></blockquote><blockquote>Run: echo ‘export STRIX_LLM=”strix/gpt-5&quot;’ &gt;&gt; ~/.bashrc &amp;&amp; source ~/.bashrc — Strix also auto-saves config to ~/.strix/cli-config.json on first successful run.</blockquote><h3>Running Your First Scan</h3><h3>Step 1 — Pull the Docker Sandbox (First Run Only)</h3><p>On your very first scan, Strix automatically downloads the sandbox Docker image. This is a one-time step, subsequent scans start immediately.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*9GXE3pYU5ikGivFbO78qwg.png" /><figcaption><em>Strix pulling the Docker sandbox image on first run</em></figcaption></figure><h3>Step 2 — Choose Your Target</h3><h3>Local Codebase</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/934/1*Zimiz6Q4bX05IiveJGi0qg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*0eTFsMYoYFJnS2ySrczbcg.png" /><figcaption><em>Strix TUI showing live scan with 4 active agents and real-time token counter</em></figcaption></figure><blockquote><strong>Token counter is live</strong></blockquote><blockquote>Notice the ‘Tokens used: 12,847’ field in the TUI header above. This counter updates in real time as agents consume API credits. It gives you full visibility into credit spend during the scan.</blockquote><h3>GitHub Repository</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/931/1*wU9zzh9t0OUqGKOOk2bl2g.png" /></figure><h3>Live Web Application (Black-Box)</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/933/1*rTtitqnWbyAoigUJNunihg.png" /></figure><h3>Authenticated Testing (Grey-Box)</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/940/1*GytlS5wTQGaf4ajJL_N_1Q.png" /></figure><h3>Multi-Target White-Box Assessment</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/934/1*u96GWCvwUL25bvQcXUBUbQ.png" /></figure><h3>Step 3 — Scan Modes</h3><p>Choose the right scan depth for your context:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/934/1*1yvlJHJDmu-2BPMOmIpayQ.png" /></figure><h3>Step 4 — Non-Interactive Mode (CI/CD)</h3><p>Use the -n flag for automated pipelines. Strix exits with a non-zero code if vulnerabilities are found, making it easy to block failing builds:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/934/1*kRqBo1VvOsl54Xx3tDItKQ.png" /></figure><h3>Step 5 — Review Your Results</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/937/1*QPSwK4Utp24As4s2MLSCqg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/972/1*hdW0QIR0tYSP-MUTgw1ugg.png" /><figcaption><em>Strix report showing 2 confirmed findings with PoC exploit and token consumption summary</em></figcaption></figure><p>Each report includes the confirmed vulnerability, its severity, the exact PoC used to validate it, and remediation guidance with specific code-level recommendations.</p><blockquote><strong>Quick Reference</strong></blockquote><h3>Common Commands</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/943/1*F1N1AicBg6PYONinachcLQ.png" /></figure><h3>Useful Links</h3><p>• Documentation: docs.strix.ai</p><p>• GitHub: github.com/usestrix/strix</p><p>• Strix Router (credits &amp; API keys): models.strix.ai</p><p>• LLM Provider Guides: docs.strix.ai/llm-providers/overview</p><p>• CI/CD Integration: docs.strix.ai/integrations/ci-cd</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f475a2bf896d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Authentication Vs. Authorization]]></title>
            <link>https://medium.com/@arinolatemilola/authentication-vs-authorization-f409cc44d5d4?source=rss-30ccf3daeca5------2</link>
            <guid isPermaLink="false">https://medium.com/p/f409cc44d5d4</guid>
            <dc:creator><![CDATA[Victoria Arinola]]></dc:creator>
            <pubDate>Thu, 07 Aug 2025 14:44:23 GMT</pubDate>
            <atom:updated>2025-08-07T14:44:23.762Z</atom:updated>
            <content:encoded><![CDATA[<p>When I began my APISEC journey, one concept I knew I had to grasp deeply was the difference between Authentication and Authorization. These two form the bedrock of secure API design, and there’s no way around it. You’ll meet them again and again, especially during every hands-on hacking session, trust me, it’s not optional!<br>So, the course for week 2 covered “API Authentication” by <a href="https://www.linkedin.com/company/apisec-university/">APIsec University</a>, and here’s where the real learning kicked off.</p><p>🔐 Authentication vs. Authorization<br>I used to confuse them (who hasn’t?), but now I clearly see:<br>Authentication = Who you are<br>Authorization = What you can do<br>Authentication comes first; it verifies your identity. <br>Authorization follows; it checks your permissions. They’re partners in crime prevention, but serve very different purposes.</p><p>A few highlights from the Course<br>🔑 Basic Authentication — simple, but insecure without HTTPS.<br>🔑 API Keys — useful, but limited when it comes to fine-grained access control.<br>🔑 Bearer Tokens — essential for client authentication in many modern APIs.<br>Then came the star of the show: OAuth 2.0 — a powerful delegation protocol that introduced me to:<br>Roles like Resource Owner, Resource Server, Client, and Authorization Server.<br>Access Tokens vs. Refresh Tokens — knowing when and how to use each is key for balancing security and user experience.</p><p>🟨 Real-world tips: Never store tokens in localStorage, Always use HTTPS, and Implement token expiration.</p><p>I also explored JWTs (JSON Web Tokens) — a favorite! Their structure (Header, Payload, and Signature) blew my mind. It’s fascinating how the signature guarantees token integrity, and how scopes and claims define the limits of access.<br>🔍 Token Introspection <br>Token introspection — this concept stood out to me, as it validates token legitimacy in real-time.<br>This course has definitely leveled up my understanding of secure API design, and I’m excited for what’s next!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f409cc44d5d4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Positioning “TRANZ” for SOC 2 Type I: A Practical Guide to Readiness and Compliance]]></title>
            <link>https://medium.com/@arinolatemilola/positioning-tranz-for-soc-2-type-i-a-practical-guide-to-readiness-and-compliance-cd0a08526767?source=rss-30ccf3daeca5------2</link>
            <guid isPermaLink="false">https://medium.com/p/cd0a08526767</guid>
            <dc:creator><![CDATA[Victoria Arinola]]></dc:creator>
            <pubDate>Tue, 29 Apr 2025 12:32:29 GMT</pubDate>
            <atom:updated>2025-04-29T12:32:29.942Z</atom:updated>
            <content:encoded><![CDATA[<p>This article details how we are structuring and positioning TRANZ (the “Company”) for SOC 2 Type 1 compliance. SOC 2 Type I compliance is a big deal for tech companies, especially ones like TRANZ that are managing sensitive customer data. Trust is becoming crucial in today’s digital environment. It is not merely a “nice-to-have” for SaaS companies to be exhibiting a strong commitment to security and data protection; rather, it is a business necessity. Customers are expecting it, and it is helping build long-term credibility in a crowded market.</p><p>This guide is breaking down how TRANZ is preparing for their audit. It is not just high-level theory — it is a step-by-step look at how we are tackling scoping, identifying gaps, fixing issues, documenting everything properly, and getting audit-ready. <br> By following these steps, you are moving confidently toward compliance, improving your company’s controls, and making customers and stakeholders feel secure.</p><p><strong>First: Understanding What SOC 2 Is Really About</strong><br> SOC 2 is based on five “Trust Services Criteria” set by the AICPA (American Institute of Certified Public Accountants). These criteria are focusing on different aspects of protecting data:</p><ul><li><strong>Security</strong>: Keeping systems protected from unauthorized access (both digital and physical).</li><li><strong>Availability</strong>: Making sure systems are up and running when customers are expecting them to be.</li><li><strong>Processing Integrity</strong>: Ensuring that systems are doing what they’re supposed to, accurately and reliably.</li><li><strong>Confidentiality</strong>: Keeping sensitive information protected according to agreements or policies.</li><li><strong>Privacy</strong>: Managing personal information properly — collecting, using, storing, and destroying it securely.</li></ul><p>For TRANZ, Security is the baseline — it applies to everyone. But because they are a SaaS company, Availability and Confidentiality are becoming really important too. So they are focusing on things like multi-factor authentication (MFA) for security, redundant systems for availability, and encryption to protect customer data.</p><p><strong>Setting Clear Boundaries (Scoping the System)</strong><br> Before going into the controls they might have in place, we are first figuring out what parts of their system are becoming part of the audit. They are using a major cloud provider, with virtual servers, databases, and networking services. In-scope systems include:</p><ul><li>Their main SaaS app (what customers are using)</li><li>APIs that connect to other services</li><li>Internal admin tools that support their operations</li><li>All customer data being processed by the platform, plus sensitive internal business data</li></ul><p>They are also creating a detailed diagram showing how everything is connected. This helps both their team and the auditors understand what’s involved and who’s responsible for what.</p><p><strong>Gap Analysis (Gap Assessment): Finding Out What’s Missing</strong><br> To figure out where they stand, we are doing a “gap analysis” for TRANZ — basically, comparing what they are already doing against what SOC 2 is expecting.</p><p>Some gaps we found:</p><ul><li>Missing multi-factor authentication (MFA) for important systems</li><li>Encryption policies that are not being fully fleshed out (especially for data at rest and in transit)</li><li>Informal incident response processes (nothing written down)</li></ul><p>We are prioritizing the gaps by risk level. Top action items include rolling out MFA, writing strong encryption standards, and formalizing how they are managing changes to systems.<br> This process is giving them a clear game plan to fix things before involving the auditors.</p><p><strong>Designing and Implementing the Controls</strong><br> After identifying gaps, we are going further to put new controls in place. Some key things we are doing:</p><ul><li><strong>Access Control</strong>: Setting up role-based access (only giving people what they need for their jobs) and enforcing MFA.</li><li><strong>Change Management</strong>: Documenting a formal process for making system changes , requiring approvals, testing, and scheduled rollouts.</li><li><strong>Incident Response</strong>: Building a plan outlining who is doing what during a security incident and training the team with drills.</li><li><strong>Monitoring</strong>: Deploying logging and alert systems to catch suspicious activity quickly.</li></ul><p>The goal is to bake security into everyday operations , not just treat it like a box to be checked off.</p><p><strong>Documenting Everything (Because the Auditor Is Asking)</strong><br> Good documentation is vital for SOC 2. We are making sure TRANZ’s key policies are clear, thorough, and up-to-date, including:</p><ul><li><strong>Information Security Policy</strong>: The overall rules that everyone follows.</li><li><strong>Data Retention Policy</strong>: How they are storing, archiving, and deleting data.</li><li><strong>Disaster Recovery Plan</strong>: How they would bounce back if systems went down.</li></ul><p>We will be using version control (like you would for software) to track policy changes, which is making it easy to show auditors how policies are evolving.<br> The trick is to write policies that are detailed enough for compliance, but simple enough for real people to actually follow.</p><p><strong>Finally: Preparing for the Audit</strong><br> Once the controls are in place and the paperwork is ready, TRANZ is turning its focus to audit prep.<br> They are choosing an auditor who knows SaaS companies inside and out.<br> They are gathering evidence — like screenshots, logs, and documents — and organizing it neatly.<br> They are doing a “pre-audit” check to catch any last-minute issues.</p><p>During the actual audit, being responsive and organized makes a big difference — it is keeping things smooth and collaborative rather than stressful.</p><p><strong>Conclusion: SOC 2 Compliance Isn’t Just About Passing an Audit</strong><br> Going through this process makes TRANZ stronger overall , not just “compliant.” It is showing customers they are taking security seriously, building trust, and setting them apart from competitors.<br> But achieving SOC 2 Type I is just the beginning. Maintaining it means keeping controls strong, regularly reviewing processes, and improving over time.<br> For any company starting this journey, the advice is: start early, involve the whole team, stay organized, and be proactive. It is hard work, but the payoff is well worth it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cd0a08526767" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>