About

I’m Jonathan Ballard, a software engineer working on AI agents and cloud infrastructure. I write about the details that need care: what an agent can access, how its tools behave, and how to test and deploy changes.

What I write about

Agent security & IAM

Scoping credentials to a task, reviewing IAM policies, and handling untrusted inputs. The articles walk through failure cases, including overly broad permissions, prompt injection, and secrets exposure.

Infrastructure as code for agent systems

Terraform workspaces, state management, and environment configuration, with examples you can adapt and test against your own requirements.

CI/CD for agent deployments

Evaluation gates, staged rollouts, and rollback plans for software that uses agents. I focus on what you can check before and after a deployment.

Background

My background is in full-stack engineering. Through Anonymous Function Labs, I offer help with cloud infrastructure, agent systems, and engineering leadership. I also build projects to explore where these tools can be useful.

Get in touch

If you have a task that takes too much manual review or a system you need help building, tell me about it. A concrete example is a useful place to start. You can also try the free IAM skill on the skills page and share what worked or what it missed.