my-agentic

A manifesto for AI-first engineering, turned into a repeatable pipeline.

This is a Claude Code plugin that carries a feature from a one-line idea to reviewed, verified, shipped code through 19 atomic, stack-agnostic skills and a TDD implementation engine — with a living roadmap above the per-feature flow.

It exists because "we use AI" and this are not the same thing. The difference is process — the discipline of a chain instead of a single prompt. That chain is what turns idea → research → requirements → architecture → design → code → tests → ship into one system where every link adds context and a checkpoint, instead of one shot from idea straight to code (and a prayer).

How the manifesto shows up in the pipeline

• AI is the first lens, not a tool bolted on. We design through AI from the first

conversation. The pipeline starts before code — with structure, not a prompt.

• The engineer stays architect and quality controller. Every skill is Socratic — it

walks decisions with you and refuses to dump a wall of output. You review, you decide.

• Domain anchor first. survey maps the repo before anything else runs, so the AI

works inside your architecture instead of guessing at it.

• Speed is risk reduction. Gated stages keep the loop tight — a stage hard-refuses when

its prerequisite artifact is missing, so mistakes surface early and cheap.