Inspiration
Managing Linux servers often means logging in via SSH, running repetitive commands, and reacting to issues after they’ve caused downtime. I wanted an AI-powered way to safely automate these tasks.
What it does
LinOps is an AI-powered Linux Operations Agent that: • Executes safe runbooks for common server tasks. • Chains multiple actions from a plain-English request. • Rejects unsafe or destructive requests. • Runs safe shell commands for requests outside its runbook library. • Automatically detects and fixes common issues like crashed services.
How we built it
Backend: Python CLI powered by Modal Functions for isolated execution in Linux containers. AI Planner: OpenAI model to map natural language requests to safe runbooks or generate safe shell commands. Safety Layer: Blocks destructive commands before execution. Auto-Heal System: Periodically checks system health and fixes detected problems.
Challenges we ran into
Designing a safety filter that prevents dangerous commands while still allowing flexibility. Getting the AI planner to correctly chain actions without hardcoding solutions. Learning Modal’s deployment and execution model under time constraints.
Accomplishments that we're proud of
Fully working AI agent that can fix a crashed service in real time without manual input. Dynamic chaining of actions from plain-English input. End-to-end automation from user request → AI plan → secure execution → JSON output.
What we learned
How to use Modal to run and manage containerized operations. How to structure a runbook library for AI-driven automation. How to integrate safety checks into autonomous command execution.
What's next for LinOps
Expand runbook library with more operational tasks. Add a web UI for monitoring and triggering actions. Integrate with cloud provider APIs for infrastructure-wide automation.
Log in or sign up for Devpost to join the conversation.