Playbooks

How I build.

The repeatable methods behind the work. Not theory, and not a list of tips. These are the systems I actually run to ship across a portfolio of live sites. Steal the thinking.

12 playbooksevery one a system I runBrowse the skills →

The system

How the methods connect

Every playbook below serves one operating loop. I set the priorities; a fleet of AI agents runs the work across the whole stack a company needs, from engineering and content to SEO, AEO, GEO, analytics, and growth.

playbooks

methods I actually run

loop stages

plan, dispatch, ship, measure

families

run the engine · shape what ships

Plan

Set priorities. Plans live in git and get reviewed like code.

Dispatch

A fleet of agents picks up the work, each isolated and right-sized.

Ship

They build and publish: software, content, and search visibility.

Measure

Status is observed, not declared. Then the loop repeats.

one loop · repeats continuously

Run the engine

6 methods

How the machine itself operates: dispatching agents, acting safely, and trusting only what it can verify.

The autonomous dispatch methodThe platform operating modelDry-run by defaultStatus is observed, not declaredWhat an agent should remember (and forget)The database is a projection of a Markdown file

Shape what ships

6 methods

What the machine produces and how it gets found: the software, the structure, and the search and answer-engine visibility.

Quality-first programmatic SEOThe content do-no-harm gateGetting cited by answer enginesOne engine, many domainsBuilding an entity graph in JSON-LDPrerendering a React SPA for crawlers

Method & PlaybookReady

The autonomous dispatch method

How I run five to ten autonomous coding agents at once, safely, cheaply, and on repeat.

Inside

The core ideaWhat not to automateFive rulesModel tieringAutonomy postureA worked exampleThe run loopCollision & isolation

9 min readRead the playbook

Framework & SystemReady

Quality-first programmatic SEO

How to publish programmatic pages that earn their place after the 2026 spam update, instead of thin pages that get demoted.

Inside

What Google targetsThe scarThe earn-the-index barSix build rulesData grain & hubsQuality-gated indexingDiagnosing drops

9 min readRead the playbook

Operating DoctrineReady

The platform operating model

Ten rules for building an autonomous platform that stays durable as it grows.

Inside

Checkable rulesThe ten rulesBuilt means exercisedPropose, do not actDeprecate before deleteWritten is not enforced

10 min readRead the playbook

Quality GateReady

The content do-no-harm gate

A grounded way to grade content for SEO, AEO, and GEO so it helps the site and never hurts it.

Inside

Help or harmThree outcomesWhat it composesWhen block firesCheapest-first upgradesAn honest gate

7 min readRead the playbook

Framework & SystemReady

Getting cited by answer engines

AEO and GEO in plain terms: how to be the answer an engine extracts and the source a generative model cites, even from a small site.

Inside

AEO vs GEOWhy the game changedWhat earns the citationWhat never gets citedCited without rankingThe levers

8 min readRead the playbook

Architecture & PatternReady

One engine, many domains

The multi-tenant pattern behind a sports network, a 50-state platform, and a family of atlases: one backend and one frontend serving many public sites.

Inside

The wrong unit of workWhat differs per tenantThe extraction is the productWhat it costsDo not extract too early

8 min readRead the playbook

Framework & SystemReady

Building an entity graph in JSON-LD

Telling search engines and language models who you are by declaring your interconnected identities, instead of hoping machines infer them.

Inside

Declare, do not inferA canonical anchorThe mesh resolves both waysMatch realityWhere it gets emitted

7 min readRead the playbook

Build Log & MethodReady

Prerendering a React SPA for crawlers

How a page that renders perfectly in a browser can serve a crawler a blank shell, and the build-time fix that makes a single-page app fully indexable.

Inside

The symptomWhy it happensStatic vs fetched bodiesThe fix: bake it inThe serving layer lies

8 min readRead the playbook

Safety PatternReady

Dry-run by default

The one rule that lets me hand real work to an autonomous system and walk away: nothing writes to the outside world until I say so, and everything it does, it can undo.

Inside

The default is show-meActing is explicitEverything has an undoWhy this is the whole game

6 min readRead the playbook

Reliability PrincipleReady

Status is observed, not declared

Why my command center refuses to trust its own status fields, and treats the gap between claimed and true state as the real signal.

Inside

The lie of the status columnReconcile, do not trustProbe the worldThe green light that lied

7 min readRead the playbook

Method & PlaybookReady

What an agent should remember (and forget)

The line between stateless agents and ones drowning in notes is ruthless curation of what actually persists across sessions.

Inside

A transcript is not a memoryFour kinds of durable factFeedback is highest valueForgetting is a feature

7 min readRead the playbook

Architecture & PatternReady

The database is a projection of a Markdown file

Why the source of truth for everything my command center does is a git-tracked text file, and the database is a view I can drop and rebuild.

Inside

Invert the truthPlans live in gitCode review on intentionsThe projection must be honest

7 min readRead the playbook

Have something you need built or fixed?

I build production Django / Next.js platforms and human-supervised AI-agent systems. Solo, senior, and fast. Tell me what you are building.

Start a project