Active · governed by a living Constitution

A 15-role AI team that builds software autonomously

Claude Corp is a multi-agent engineering system built on Anthropic's Claude. Specialized AI roles plan, build, test, review, and deploy across a portfolio of projects — all governed by a living Constitution.

15
Specialized roles
7
Design-Compiler layers
5
Projects · 2 fully live
v1.12
Living Constitution

The orchestration pipeline

Sky
sets intent
a directive
Morgan
plans & routes
dependency graph
15 Roles
build in parallel
scoped branches
Safety
7-layer compiler
a11y · privacy · perf
main
Sky merges
human approval

Select a stage to follow a task from intent to merge.

Proof, not promises

A real system, with receipts

Extraordinary claims deserve evidence. Everything here is verifiable — open the live apps, inspect the source, and read how the system is governed.

Versioned governance

  • v1.12 Living Constitution — Art. 17 merge gate
  • OS Agent OS runtime overlay, versioned per release
  • order Sky → Constitution → roles → skills

Governance is dated and citable — not vibes. No agent self-amends; main is sacred and only Sky merges.

Read the model on GitHub (opens in new tab)

Every role, its own branch

  • feat/claudecorp-p1-glass-foundation
  • attribution/claude-corp-truth-2026-06-18
  • content/portfolio-voice-pass-2026-06-10
  • perf/auto-2026-05-31

Real, role-prefixed branches across projects. Agents propose on scoped branches; nothing reaches main without Sky's review.

This page's own 6-phase build history (opens in new tab)

Continuously audited

900+
QA & audit reports written across the portfolio.

Multi-agent reviews truth-audit each other against live code and git history before anything is called done.

How it works

From backlog to deploy, automated

Each role is a Claude agent with a scoped skill file. Morgan orchestrates across roles; the Constitution enforces every safety gate.

01

Sky sets intent

Sky writes a directive or runs a slash command. The authority order locks in: Sky > Constitution > role files > skills.

02

Orchestrator plans

Morgan reads backlogs, open branches, and QA reports — then routes work to the right specialists in dependency order.

03

Roles execute

Each agent works in isolation on a scoped branch. Design, code, test, a11y, performance, and privacy checks all run before merge.

04

Safety sweep

A 7-layer Design Compiler gate plus the Orion recovery agent ensure no UI ships without full parity and a rollback path.

05

Sky merges

Only Sky touches main. Agents branch, propose, and report — never auto-merge. Morgan briefs; Sky decides.

The team

15 specialized roles

Each role is a purpose-built Claude agent with a scoped skill file, a prefixed branch namespace, and a silent external presence — only Morgan talks to Sky.

Select any role to see its authority, limits, and constraints.

Design & Frontend Backend & Data Safety, Quality & Infra Product, Research & Policy Coordination
M
Morgan
Project Manager
Status briefings, decision routing, and the sole channel for messaging Sky.
S
Shamus
Feature Engineer
Ships new screens, flows, and UI. Every PR passes the 7-layer Design Compiler gate.
D
Dani
Creative Director
Owns the token-first design system and runs the Design Compiler — POLISH and BLOCK calls are hers.
A
Alex
Accessibility Engineer
WCAG 2.2 AA audits on every UI change. No screen ships without an a11y pass.
D
Dana
Backend Engineer
Schema, migrations, RLS, and data integrity. Never touches the live database.
P
Peter
Performance Engineer
Bundle audits, render profiling, and query optimization across every project.
S
Steve
Safety Engineer
Enforces Constitution safety rules. Raises blockers before risky code ships.
G
Gary
QA Engineer
Writes and runs the test suite. Catches regressions before they reach review.
R
Rory
DevOps Engineer
CI/CD pipelines, EAS builds, and release readiness. Gets things actually shipped.
Q
Quinn
Product Manager
Grooms the backlog, specs features, and prioritizes by real user value.
R
Riley
User Researcher
Personas, journey maps, and friction analysis grounded in real disabled-user needs.
J
Jordan
Privacy Advisor
Privacy-by-design gate on all location, auth, and user-data changes.
C
Casey
Community Manager
Contributor docs, onboarding, and growth strategy for the project.
W
Will
Technical Writer
READMEs, setup guides, changelogs, and the learnings log — clear language everywhere.
O
Orion
Recovery Agent
Recovery-only (Art. 8.4). Steps in when another role hits a blocker or inconsistent state.
What makes it work

Built-in guardrails, not guidelines

Claude Corp is governed by a Constitution that outranks any role or skill. Safety is structural, not aspirational.

Living Constitution

A versioned governance document (currently v1.12) that all 15 roles inherit as hard law. No agent self-amends — only Sky changes it.

7-Layer Design Compiler

Every UI change passes tokenization, accessibility parity, component consistency, visual entropy, luxury score, regression safety, and the compile decision before it's marked done.

Background Mode

Audit agents can run autonomously on a schedule, bounded by Background Mode — a halt-sentinel file freezes all background work instantly, and every cycle is capped at one reversible change.

main is sacred

No agent ever touches main. Every change lands on a role-prefixed branch, gets reviewed, and merges only when Sky approves.

One voice to Sky

Morgan is the only agent allowed to message Sky. Every other role surfaces findings to QA reports; Morgan picks them up and briefs.

Velocity Loop

Five chained commands drive a continuous backlog-to-PR cycle: plan → work → review → state-sync → next. No idle time between iterations.

Built by one developer, powered by fifteen agents

Claude Corp is the multi-agent engineering system Sky Halisky designed and runs on Anthropic's Claude — always on, always scoped, always human-approved. The software it maintains is real, live, and linked above.

Sky is open to thoughtful product & implementation collaborations.