Project Update: CyberShield AI Agent is Live! We are thrilled to post our official development log for CyberShield AI Agent! Here is a breakdown of how the project evolved from an idea to a fully functioning autonomous security copilot: Phase 1: The Pipeline Setup Connected GCP Cloud Logging with the Dynatrace MCP Server using Dynatrace Query Language (DQL) to catch live threats instantly. Phase 2: Integrating Gemini Pro Fed the live DQL streams into Gemini Pro, training the model to read context and separate true Indicators of Compromise (IoC) from normal background noise. Phase 3: Automated Remediation Loop Coded the feedback loop to automatically generate Google Cloud Armor WAF rules and strict Kubernetes Network Policies to isolate workloads in under 15 seconds.
Log in or sign up for Devpost to join the conversation.