The latest articles on DEV Community by Deepika bhuvaneshwaran (@deepika_bhuvaneshwaran_63). https://dev.to/deepika_bhuvaneshwaran_63 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4202017%2Fad3f246a-5221-4194-bd00-4616643d9730.png https://dev.to/deepika_bhuvaneshwaran_63 en Deepika bhuvaneshwaran Fri, 23 Oct 2026 08:27:43 +0000 https://dev.to/deepika_bhuvaneshwaran_63/building-chain-of-custody-exploring-ai-agent-safety-with-python-and-fastapi-3jo5 https://dev.to/deepika_bhuvaneshwaran_63/building-chain-of-custody-exploring-ai-agent-safety-with-python-and-fastapi-3jo5 <h1> Building Chain of Custody: Exploring AI Agent Safety with Python and FastAPI </h1> <p>As AI agents become more capable, one important question comes to mind:</p> <p><strong>How can we make AI agent actions safer, more transparent, and easier to review?</strong></p> <p>As a Computer Science and Engineering student interested in AI agents and cloud technologies, I explored this question by building a project called <strong>Chain of Custody — AI Agent Safety and Decision Monitoring</strong>.</p> <h2> 🚀 What is Chain of Custody? </h2> <p>Chain of Custody is a web-based dashboard designed to help users review proposed AI agent actions, evaluate potential risks, and inspect safety decisions.</p> <p>The project focuses on three possible outcomes:</p> <ul> <li> <strong>PROCEED:</strong> Allow an action when it meets the applicable safety checks.</li> <li> <strong>CONFIRM:</strong> Require additional human review before proceeding.</li> <li> <strong>DENY:</strong> Block an action that fails the applicable safety checks.</li> </ul> <p>These outcomes help demonstrate how a safety-review workflow can make agent actions easier to understand.</p> <h2> 🛠️ Technology Stack </h2> <p>I used the following technologies to build the application:</p> <ul> <li> <strong>Python</strong> for backend logic.</li> <li> <strong>FastAPI</strong> for API endpoints.</li> <li> <strong>HTML, CSS, and JavaScript</strong> for the web interface.</li> <li> <strong>Strands Decider integration</strong> to explore AI-assisted action evaluation.</li> <li> <strong>Vercel</strong> for deployment.</li> <li> <strong>GitHub</strong> for source control.</li> </ul> <h2> 🖥️ Key Features </h2> <h3> 1. Dashboard </h3> <p>Provides an overview of the application's safety-monitoring workflow.</p> <h3> 2. Inbox </h3> <p>Provides an interface for reviewing incoming requests.</p> <h3> 3. Agent Run </h3> <p>Allows proposed actions to be evaluated and displays a PROCEED, CONFIRM, or DENY outcome.</p> <h3> 4. Flight Recorder </h3> <p>Provides an interface for inspecting decision records and improving visibility into the review process.</p> <h2> 🧠 What I Learned </h2> <p>Building this project helped me explore how a backend API and frontend dashboard work together. It also gave me practical experience with deployment, API testing, GitHub, and designing interfaces for AI safety workflows.</p> <p>One important lesson is that an AI-generated decision should not automatically be treated as a guarantee of safety. Real-world systems need appropriate authorization checks, human oversight, testing, and reliable audit records.</p> <h2> 🔭 What's Next? </h2> <p>I plan to improve production testing, verify the hosted Decider connection through the deployed application, and explore Amazon Bedrock integration.</p> <p>These improvements will help me develop a more complete understanding of building and deploying responsible AI applications.</p> <h2> 🔗 Try the Project </h2> <p><strong>Live demo:</strong> <a href="https://chain-of-custody.vercel.app/" rel="noopener noreferrer">https://chain-of-custody.vercel.app/</a></p> <p><strong>Source code:</strong> <a href="https://github.com/Deepika20053/chain-of-custody" rel="noopener noreferrer">https://github.com/Deepika20053/chain-of-custody</a></p> <p>I'm continuing to learn about AI agents, cloud technologies, and responsible software development. I'd love to hear how other developers approach safety and oversight in agentic applications.</p> <h1> AWS #AIAgents #Python #FastAPI #AISafety </h1> ai backend python security