I'm an application modernisation specialist who fell into the world of coding agents. Now I'm documenting the journey — testing Kiro, Claude Code, Codex, and OSS agents like Pi-Agent and OpenCode, running them on local LLMs and AWS Bedrock.
Feed The Agent is a best-practices guide for teams working with AI agents — not marketing fluff, but honest notes on what works, what breaks, and how to make these tools useful at team scale.
Currently writing: Coding Agents in Teams — a series on integrating autonomous coding agents into real engineering workflows.