From empty domain to a ranking medical site, end to end

A brand-new multi-location practice, an empty domain, and all 25 MedAuthority skills in one Claude Code project. This is the entire pipeline, from market research to the first ranking report, run as a single connected workflow.

The smaller bundles each solve one job. The Full Stack is for the agency that runs the whole pipeline, and the reason it costs less than buying every bundle separately is that the bundles were always meant to chain. This post takes a single client all the way through, and the through-line matters: every skill in the sequence reads the output of the one before it, so by the end you are not looking at 25 disconnected deliverables. You are looking at one site that was planned, built, optimized, and measured as a unit.

The client is a new orthopedics group opening three locations in a single metro. Nothing exists yet: no site, no content, no profiles. We are using the $299 Full Stack, which is all 25 skills plus the START-HERE guide, the WORKFLOWS chains, and the SOP templates. We will follow the five categories in order, because that order is the workflow.

The whole stack in one project

With the Full Stack you are not importing five files, you are importing the operating system for the engagement. The START-HERE guide explains the loading pattern for each tool; in Claude Code it is the familiar CLAUDE.md import.

mkdir -p ortho-launch/.claude/skills
cp medauthority-full-stack/skills/*.md ortho-launch/.claude/skills/
cp medauthority-full-stack/WORKFLOWS.md  ortho-launch/
cp medauthority-full-stack/START-HERE.md ortho-launch/
cd ortho-launch
echo "@.claude/skills/" >> CLAUDE.md   # import the whole skills folder
echo "@WORKFLOWS.md"     >> CLAUDE.md   # the chains and bundle map
claude

WORKFLOWS.md is worth reading before you start. It is the map of which skill feeds which, so instead of guessing the order you follow the documented chains. The whole rest of this post is one of those chains, run end to end.

Strategy and discovery: decide what to build before building it

The most expensive mistake in a new-site build is building the wrong thing well. So nothing gets written until the strategy skills have run. Market Research takes the specialty, the metro, and the known competitors, and returns the shape of the opportunity: which procedures have real local demand, where the market is saturated and where it is soft, and the clusters of patient intent worth targeting. It tells you the uncomfortable truth when a segment is too crowded to win quickly, rather than the cheerful answer, because planning a site around a losing segment wastes the whole engagement.

Competitor Reverse-Engineering then takes the named competitors and maps how they actually rank: their page structure, their directory presence, the content depth that is holding their positions. Its output is a structured gap analysis. Keyword Intent Clustering takes both of those and organizes the target terms by intent, separating the "find a surgeon now" transactional searches from the "is this surgery worth it" informational ones, because those become different pages with different jobs.

Why this category feeds everything downstream

The clusters are not a deliverable you hand the client. They are the input to the sitemap. The gap analysis is not a report that sits in a folder. It is the brief for the content engine. This is the moment the chain is established: the strategy outputs are structured specifically so the architecture skills can consume them without you re-deriving anything.

Architecture: turn the strategy into a structure

The Sitemap Generator reads the keyword clusters and the gap analysis and proposes the site's skeleton: the procedure-page silos, the three location pages, the conditions content, the bio pages. It is not a generic template; it is a structure shaped by what the research found worth ranking for. Because a multi-location practice lives or dies on local relevance, the location pages are first-class, one per location, each genuinely about that location rather than a spun copy.

Internal Linking Strategist takes that sitemap and defines the authority flow: how the procedure pages support each other, how each location page connects to the services offered there, how the informational content links up to the transactional pages so the equity concentrates where conversions happen. Local Page Generation builds out the three location pages with real local specificity. The UX Audit reviews the planned structure for the trust signals a medical site needs (credentials, real provider information, an unfriction-ed path to booking) before a single content page is written.

Content operations: the highest-value cluster in the stack

This is where most of the billable value is, and where the compliance discipline earns its place. The content skills do not improvise. The EEAT Content Writer takes a page from the sitemap and writes it to be authoritative and patient-credible, hedging clinical claims rather than stating them as fact, and flagging every claim a clinician must verify. It refuses to fabricate statistics or credentials; where a real number belongs it marks the spot and asks the practice to supply the verified figure.

The Procedure Page Generator handles the workhorse pages, what the procedure is, what to expect, candidacy, the process, an honest account of benefits and risks, in a consistent structure across every procedure so the site feels like one coherent practice. The Symptom-to-Service Mapper builds the bridge between how patients search (by symptom, by problem) and the services that address them, without diagnosing the reader. The FAQ Schema Builder produces valid FAQPage JSON-LD for the voice-search and rich-result opportunity.

The gate that makes it safe

Then everything routes through the Readability and Compliance Rewriter. This is the human-review gate made into a step. It runs a plain-language pass, inserts the required disclaimers, checks for claims that crossed a line, and produces a clean draft with a list of exactly what a licensed provider needs to sign off on. Nothing from this category publishes without that review. The skills make the review fast by flagging precisely what to check; they never replace it. In a YMYL vertical that gate is not optional, and building it into the workflow as a discrete step is what lets an agency produce medical content at volume without producing liability at volume.

Local SEO: three profiles, not one

A multi-location practice has three Google Business Profiles, and they have to be consistent with the site and with each other. GBP Service Optimization structures each profile's categories and services to match the procedure vocabulary the content engine already established, so the profile and the site describe the same practice in the same words. GBP Post Generation produces educational posts per location. Review Response and Reputation Pattern Analysis handle the incoming reviews with the same HIPAA-safe discipline as the content, never confirming a person was a patient in a public reply.

The reason this comes after content and not before is that the profiles inherit the content's vocabulary. Run local first and you are optimizing profiles against services that have not been defined yet. Run it here and everything lines up.

Analytics and operations: set the baseline and make it repeatable

On a brand-new site the analytics skills play a specific role: they establish the baseline you will measure against later. The SEO Equity Analyzer records the starting position, near-total dependency on paid for now, so that next quarter's report can show the organic equity beginning to build. Lead Attribution gets wired up at launch, including call tracking, so that no booked patient is invisible from day one and you are not trying to reconstruct attribution months later.

The SEO Audit Assistant documents the technical launch state as the clean baseline. The Content Decay Detector has nothing to detect yet, but it is scheduled. And the SOP Builder captures this entire pipeline, all five categories, as the agency's standard new-practice launch procedure, so the second client's launch is a documented process rather than a heroic effort. The SOP templates in the bundle are the starting point; the builder adapts them to exactly the chain you just ran.

What the full pipeline produced

Start to finish: a market-validated strategy, a site architecture justified by that strategy, a full set of compliant medical content built on that architecture, three local profiles consistent with that content, and an analytics baseline wired to measure all of it, plus the SOPs to do it again. Twenty-five skills, but one workflow, because each skill's structured output was the next skill's input. That is the entire argument for a skill stack over a prompt pack, demonstrated on a single client: a prompt pack gives you 25 islands, and a stack gives you a pipeline.

It is also why the Full Stack is priced below the sum of the bundles. The bundles were always parts of this one machine. Buying them together and running them as the chain they were designed to be is the point.