Begin to End How Yherda actually works

Here's exactly what
happens when you start.

From zero to a Claude that knows your team, your process, and your constraints. Step by step. Nothing skipped.

Already convinced?  Skip to get started →
1

Install the agile skills into Claude

You start by giving Claude the agile skill set — a package of roles and ceremonies that Claude already knows how to run. From that point forward, Claude has the context it needs to guide you through your work in an agile way.

▼ Not familiar with agile?

Agile in 60 seconds

Agile is a way of working in short cycles — usually two weeks — where the team checks in regularly, adjusts based on what they learn, and keeps improving. Instead of planning everything up front and hoping it works out, you plan a little, do a little, learn, adjust.

The key roles: a Scrum Master who facilitates the process, a Product Owner who represents what the team should be building, and the team doing the work. The key ceremonies: planning, daily standups, a review, and a retrospective. All of that is already built into the skills you're about to install.

2

Kick off /scrum-master

If you know agile, you already recognize this role. Type /scrum-master and the skill activates. The Scrum Master already knows how to get you started — it walks you through the setup arc we've designed, asking the questions it needs in the right order.

/scrum-master
3

Answer the setup questions — Claude stores everything as beliefs

The Scrum Master asks you about your setup. Do you use Jira? Something else? How does your team work? What ceremonies do you run? What does "done" mean to you?

Every answer gets stored as a belief in your Yherda model — not just as a setting, but as something Claude understands and can reason about. "This team uses Jira and considers a story done when it's deployed and monitored" is a belief. Claude carries it into every future conversation.

4

Claude connects to your tools

Once it knows what you use, Claude connects up. Jira integration, your repo, whatever's relevant to your setup. The Yherda interface is already wired in — Claude has been talking to it from the moment the skills were installed. Now it starts using that connection to store and retrieve your team's belief state.

5

Your bootstrap belief set is already loaded

You don't start from a blank slate. We've pre-loaded a set of foundational agile beliefs — the things that any team starting this journey needs to hold before the ceremonies make sense. Here's what's already in there:

Agile Foundation
Short cycles with regular inspection beats long plans with delayed feedback.
Done Means Done
Work isn't done when it's built. It's done when it's shipped, monitored, and confirmed to behave as intended.
Retro as Learning Loop
Every sprint produces learning. That learning has to enter the system — not stay in someone's head — or it disappears.
AI as Team Member
Claude participates in ceremonies the same way a team member does — it needs the same context, the same constraints, the same goals.
Belief Sync is the Primitive
The ceremonies aren't the point. Shared belief is the point. The ceremonies are the mechanism that produces it.
6

From here: continuous improvement, forever

Every sprint, the retro updates your belief system. Every new constraint, decision, or learning goes in. Every Claude identity on the team picks it up at the start of the next session.

The system compounds. Month three is qualitatively different from week one. Not because Claude got smarter — because your belief system got richer. The context Claude works from is more accurate, more specific, more yours.

This is the loop. It never stops. That's the point.

Why does this have to be free?

A belief system that compounds over time, that carries organizational learning forward, that makes AI work from the same context as the humans around it — that can't be something only well-funded teams can afford. The reason goes deeper than pricing strategy.