The engineering log.
Notes from building developer tools, platforms and agent workflows in public: what we tried, what broke, and the guardrails we wrote afterwards.
-
Architecting Your Codebase for Modern Agentic Development
Giving an LLM a terminal and a prompt full of rules is a recipe for broken trunk branches. Here is how we architected vitruvian-core for autonomous AI coding agents: specialized subagent roles, per-agent GitHub App identities, hermetic Bazel sandboxing, fail-closed guardrails, and asynchronous watchdog loops.
Read the post -
Every AI agent gets its own GitHub identity
Ten coding agents work in our monorepo. Until recently, GitHub saw all of them as one person, so none of them could review another's work. Here is how we gave each agent its own GitHub App, what we had to measure because the docs don't say, and why identity alone...
Read the post -
The Reality of Agentic Coding Architecture
The era of “chat wrappers” is completely dead. As of Q2 2026, the transition from basic text-generation LLMs to fully autonomous agentic code workflows is accelerating rapidly. However, a stark divide is emerging between teams trying to prompt their way to success and those engineering rigorous, deterministic architectural patterns.
Read the post -
Introducing ADK Agents: Your AI-Powered DevOps Companion Archived
Meet ADK Agents - the intelligent DevOps assistant that understands your codebase, executes commands safely, and helps you ship better software faster. Built on Google's ADK framework with developer productivity at its core.
Read the post
Subscribe with RSS.