Stop Overloading Your AGENTS.md
Why repository-level agent instructions should stay lean, act as indexes, and avoid forcing every agent and user into one workflow.
Agents and Markdown
Product notes and field reports for human review.
Artifact-backed product notes alongside evergreen guides for reviewing agent-generated plans, documents, and preview applications.
Featured artifactposts/who-verifies-the-agents.md
Who Verifies the Agents?
How agents can verify AI-generated code through layered checks, deployment, end-to-end testing, and exploratory use.
Latest
Recent articles.
Nothing Stops the Train: Engineering Systems in the Age of Cheap Execution
A model for enterprise engineering when AI-assisted execution is cheap, reviews run alongside development, and verification replaces human gates.
Developing Agents: From Theory to Practice
Practical patterns for building reliable agent packs with clear roles, context boundaries, quality gates, and software-style development discipline.
Agent-Assisted PM Lifecycle
A lifecycle model for using agents across PM work from ideation and specification through acceptance testing, documentation, and release artifacts.
The AI-First PM Toolchain
Why PM artifacts should move to markdown, Git, and repository-aware AI assistance so agents can reliably read, write, and review them.