Cloud Zombie Hunter: An AI Agent vs. the Dead Weight on My Bill

Discover how an AI agent finds and quantifies cloud waste on AWS and Databricks, ensuring you only pay for what you use.

Claude Code AWS AWS CLI AWS MCP Server Databricks Unity Catalog

Overview

I built a terminal-based agent — Claude Code running locally — that hunts two kinds of cloud waste across AWS and Databricks in one session, with evidence and grounded dollar estimates, and never touches anything without my say-so.

It looks for two things:

Zombies — abandoned resources still billing. On AWS: unattached EBS volumes, unassociated Elastic IPs, NAT gateways with near-zero traffic, stale snapshots, and the one everyone forgets — orphaned CloudWatch alarms left behind when a resource is decommissioned. On Databricks: all-purpose clusters left running with no auto-termination, SQL warehouses with no auto-stop, and scheduled jobs that fail silently but keep getting provisioned.
Over-provisioned / mis-configured resources. ECS services whose CPU/memory reservation dwarfs actual utilization, or with autoscaling that never triggers (min capacity too high, scale-in disabled, or tracking a metric that never moves). Oversized EC2/RDS vs Compute Optimizer. Databricks clusters with oversized worker types, autoscaling minimums set too high, all-purpose clusters doing jobs that belong on job clusters, or warehouses sized XXL where LARGE would do.

Tech stack

Claude Code

Anthropic's agentic coding tool: Unleash Claude's raw power directly in your terminal or IDE to turn complex, hours-long workflows into a single command.

Claude Code is Anthropic’s powerful agentic coding assistant, designed for high-velocity development. It operates natively within your terminal, IDE (VS Code, JetBrains), or via a web interface, allowing you to delegate complex tasks like feature building, bug fixing, and codebase navigation. The agent plans, edits files, executes commands, and creates commits, maintaining awareness of your entire project structure. Internally, Anthropic engineers using Claude Code reported a 67% increase in productivity, demonstrating its capacity to deliver significant gains for Pro and Max plan users.

https://claude.com/code

View projects
AWS

AWS is the world's most comprehensive and broadly adopted cloud platform, offering over 200 fully featured services from 33 geographic Regions.

AWS is the global leader in cloud infrastructure, delivering over 200 fully featured services. We operate across 105 Availability Zones within 33 geographic Regions, ensuring high availability and low latency for your applications. Core services like Amazon EC2 (virtual servers), Amazon S3 (scalable object storage), and AWS Lambda (serverless compute) provide the foundational building blocks for any workload. This platform allows customers (from startups to Fortune 500s) to innovate faster, reduce operational costs by moving from CapEx to OpEx, and scale instantly. Security remains paramount: we offer 300+ security, compliance, and governance services, meeting standards like ISO 27001 and SOC 1/2/3. Simply put, AWS provides the secure, flexible, and proven foundation you need to build anything.

https://aws.amazon.com/

View projects
AWS CLI

A unified command-line tool to control, automate, and manage all your Amazon Web Services from a single terminal interface.

The AWS Command Line Interface (AWS CLI) is the go-to tool for developers and operators looking to bypass the AWS Management Console and control their cloud infrastructure directly from the terminal. By installing a single tool, you gain scriptable access to virtually every AWS service (from provisioning Amazon S3 buckets to launching Amazon EC2 instances). It simplifies complex workflows through powerful features like multi-part file uploads, automatic pagination, and output formatting in JSON, YAML, or text. Whether you are running quick ad-hoc queries or building automated CI/CD pipelines, the AWS CLI provides the speed and precision needed to manage your cloud resources efficiently.

https://aws.amazon.com/cli/

View projects
AWS MCP Server

The AWS MCP Server gives AI coding agents secure, authenticated access to more than 15,000 AWS API operations and live documentation through a compact, standardized toolset.

The AWS MCP Server (part of the Agent Toolkit for AWS) acts as a managed bridge between your AI assistants and your cloud infrastructure. By implementing Anthropic's Model Context Protocol, it exposes a minimal tool footprint (including call_aws and search_documentation) to prevent model context bloat while granting agents the ability to safely query docs, execute API calls, and run sandboxed Python scripts. Security remains tight: the server relies on standard AWS SigV4 signing, integrates with IAM and Service Control Policies to enforce strict read-only or mutating boundaries, and requires zero configuration changes when AWS rolls out new APIs.

https://github.com/awslabs/mcp

View projects
Databricks Unity Catalog

A unified, open governance solution for managing data, analytics, and AI assets across multiple clouds.

Databricks Unity Catalog acts as a centralized governance layer that secures and audits files, tables, machine learning models, and dashboards. It structures data organization through a clean three-level namespace (catalog.schema.table) and enforces consistent access policies across workspaces. With features like automated data lineage, attribute-based access control, and built-in data quality monitoring, it simplifies compliance and discovery (allowing teams to track data flow from raw storage to final BI reports) while operating on open-source standards.

https://www.databricks.com/product/unity-catalog

View projects

Cloud Zombie Hunter: An AI Agent vs. the Dead Weight on My Bill

Related talks