🏠 Leadership Team

AI boardroom for independent property agents

Claude Code Multi-Agent Skills + Subagents HTML / JS Markdown State

📋 Project Overview & Problem Statement

Challenge: An independent property agent is a one-person agency — chasing listings, writing copy, running open houses, following up buyers, negotiating offers, and trying to grow a personal brand at the same time. Leads slip through WhatsApp, listings sit too long, and the principal agent has no one to strategise with on a Monday morning.

Solution: Leadership Team is a 7-seat autonomous boardroom powered by specialised Claude Code subagents. Each Monday, the principal agent runs run board meeting and six AI managers deliver reports on listings, leads, content, automation, comparables, and strategy — all grounded in a living company-state.md that tracks every listing, buyer, and deal. The visual boardroom (Zoom-style grid) gives the meeting a face — a tactile mockup of the leadership team behind the one-person agency.

Key Benefits for Property Agents

🖥️ Application Features

🎥 Zoom-Style Boardroom

A 7-seat video grid with editable names, uploadable photos, active-speaker highlight, and a live meeting timer. Designed to make the weekly listing review feel real.

📋 Listing-Driven Agenda

A side panel with click-to-edit agenda items that walks through hot listings, warm leads, and pending offers each week. Persists in the browser.

🗣️ run board meeting

A single command inside Claude Code orchestrates Siti drafting the agenda, each manager reporting on their corner of the agency, Mei Ling observing, and minutes written automatically to disk.

✅ check approvals

A 5-minute daily review of pending approval.md. Price cuts, counter-offers, co-broke requests, and ad spend are surfaced with context so the agent can approve, defer, or reject in one pass.

🤖 AI Integration & Intelligence

👥 Seven AI Roles Tailored to a Property Agency

Faizal Tan (Engineer — WhatsApp + portal automation), Rajesh Kumar (Research — comparable sales, market movement), Jason Lim (Marketing — lead pipeline + open houses), Anisa Rahman (Editor — listing copy, reels, captions), Siti Nurhaliza (Coordinator — viewings + minutes), Mei Ling (Mentor — positioning & deal strategy), and Amir Hakim (Principal Agent).

🧠 Grounded in Agency State

Every manager reads company-state.md — the live listing roster, buyer pipeline, and deal log — before reporting, so their contributions reflect the actual week, not hallucinated context.

📈 Rajesh — Market Research

Pulls comparable sales for new mandates, tracks median price-per-square-foot in the agent's target zones, and flags when a listing's days-on-market crosses a warning threshold.

🎓 Mei Ling — Deal Strategy Mentor

Offers a cross-cutting perspective on hard calls — accept a co-broke, hold the price, invest in paid ads, or hire an admin — during meetings and in private 1:1 sessions between meetings.

🛠️ Technical Architecture & Implementation

Frontend (Boardroom UI)

HTML5 CSS Grid Vanilla JS localStorage

Agent Layer

Claude Code Subagents Skills Slash Commands

State & Storage

Markdown Files Local Folder Git (optional)

System Architecture

📖 How to Run

Prerequisites

Weekly Meeting

# Inside a Claude Code session run board meeting # Siti drafts the agenda # Each manager reports in turn # Mei Ling observes and gives strategic notes # Minutes written to what was discussed.md

Daily Approval Review

check approvals # Reads pending approval.md # Surfaces each item with context # Founder approves, defers, or rejects # Updates state + notifies affected managers

Open the Visual Boardroom

# Double-click the file, or open in any browser boardroom.html

📊 Key Metrics

7
Boardroom Seats
6
AI Manager Agents
2
Core Commands
0
Backend Services

Business Value for Property Agents