AI Founder OS
Reference

Docs

Three documents that describe how this system works, how to debug it, and where it breaks under load. Read them before writing task scripts. Reference them inside task scripts. They exist so you and Claude are working from the same facts.

SSOT feeding hotspots with a no drive-by edits warningSingle Sourceof TruthHotspotspage.tsxroute.tscomponent.tsxNo drive-by edits outside hotspots

How docs connect to code

Each document describes a part of the single source of truth.

1

Architecture Blueprint

What it is

A complete map of every folder, route, data flow, and auth path in the repo. Covers deployment model, environment variable tiers, and the exact rules that prevent architectural drift.

When to use it

Read it before your first feature commit. Return to it whenever you are unsure where a new file, route, or dependency belongs.

How to use it in the operator loop

Before you write a task script, open the blueprint and copy the relevant folder responsibilities and invariants into the task. This prevents Claude from inventing new patterns that conflict with the existing structure.

2

Debug Protocol

What it is

A step-by-step procedure for diagnosing build failures, runtime errors, and deployment issues. Includes a triage checklist, root cause categories, and a template for documenting findings.

When to use it

Use it the moment a build breaks or a page returns an unexpected error. Follow the steps in order instead of guessing.

How to use it in the operator loop

Paste the relevant triage checklist into your Claude task as a constraint. Tell Claude to follow the debug protocol before making any fix, and to record findings in the format the document specifies.

3

Scale Simulation

What it is

A structured walkthrough of what happens to this architecture under real load. Identifies which parts survive, which parts break, and at what thresholds you need to act.

When to use it

Review it before making infrastructure decisions or adding features that touch auth, data storage, or API routes. It tells you what is safe to defer and what is not.

How to use it in the operator loop

When planning a task that touches a known scaling boundary (Clerk metadata, webhook reliability, rate limits), reference the simulation thresholds in your task constraints so Claude stays within safe limits.

Operator Assistant

Not sure which document applies?

Paste your error, your build log, or your question. The assistant will reference the right doc and surface the relevant section.

Want execution help?

The Operator Assistant, Activator scripts, and prompt library are available to active members.

These documents describe what exists, not what should exist. Update them when the architecture changes. If a document and the code disagree, the code is right and the document needs a fix.