Give Your Org a Spark ✒️⚡

Build a Chief Agentic Officer — a multi-agent AI system that runs your organization's operations.

8 AI agents. Shared state. Persistent memory. Governance built in. $20/month.

Spark is the operating system behind Radical Imagination, a cultural heritage technology collective managing 17 projects across 154K+ lines of code. It handles grants, curricula, cultural validation, impact reporting, outreach, and portfolio tracking — all through a multi-agent architecture running on Claude Code.

This repo is the template. Fork it. Customize it. Give your org a Spark.

---

What You Get

├── CLAUDE.md                    # Your org's instructions for Claude Code
├── QUICKSTART.md                # 15-minute setup guide
├── GOVERNANCE.md                # Operational governance model
├── state/
│   ├── org-state.md             # Priorities, active projects, deadlines
│   ├── portfolio.md             # Every project with status
│   ├── last-session.md          # 15-line quick-start for every session
│   ├── decision-journal.md      # Every decision with reasoning
│   └── roster.md                # People, roles, connections
├── agents/
│   ├── router.md                # How Spark routes to the right agent
│   ├── grant-writer.md          # Sample: grant strategy agent
│   ├── content-creator.md       # Sample: content/social agent
│   └── reviewer.md              # Sample: QA/validation agent
├── memory/
│   ├── MEMORY.md                # Cross-session memory index
│   └── example-feedback.md      # Sample memory entry
├── protocol/
│   ├── session-start.md         # What happens at the start of every session
│   └── session-close.md         # What happens at the end of every session
├── docs/
│   ├── agent-spec-standard.md   # How to define agents (the portable format)
│   ├── open-infrastructure.md   # How .claude/ works as an OS
│   ├── positioning.md           # How Spark differs from CrewAI, LangGraph, AutoGPT
│   ├── honest-failures.md       # What went wrong and what the system learned
│   └── case-study-dubois-does-data.md  # One project through all 8 agents
└── examples/
    ├── money-agent-grant-strategy.md   # Real grant strategy output
    ├── validator-review.md             # Real cultural validation review
    └── cross-project-scan.md           # Real portfolio-wide intelligence scan

What Spark Does (at Radical Imagination)

Agent	What It Does	What It Replaces
Curriculum Architect	Designs learning experiences	Instructional designer ($65-85K)
Money Agent	Grant strategy, funder analysis	Grant writer ($50-70K)
Cultural Archivist	Source registries, provenance	Research archivist ($55-75K)
Cultural Validator	Reviews for cultural integrity	Cultural consultant ($100-250/hr)
Technical Mentor	Code guidance, learning paths	Senior dev mentor
Collective Manager	Roster, portfolio, onboarding	Operations manager
Impact & Engagement	Reports, social, content	Impact analyst ($30-40K)
Commercial Strategist	Evaluates opportunities	Business dev

What these agents share: the same state files, the same memory, the same governance protocols.

How It Works

1. You talk to Spark. One entry point, not 8 separate tools.
2. Spark routes to the right agent based on what you're asking.
3. Every agent reads the same state files — your portfolio, priorities, people, decisions.
4. Every agent writes to the same state files — the institution documents itself.
5. Memory persists across sessions — Spark remembers your preferences, your corrections, your context.
6. Every session starts and ends with a protocol — read state, check for contradictions, do the work, update state, draft the morning email.

The state files ARE the institution. If the founder gets hit by a bus, the state files tell the next person everything.

Quick Start

Prerequisites

• Claude Code subscription (~$20/month)
• A text editor (VS Code, Cursor, or any)
• A Gmail account (for email drafting via MCP)
• A project or organization you want to build Spark for

Setup (15 minutes)

# 1. Fork this repo
gh repo fork ShdwSpde/spark-workshop-template --clone

# 2. Open in your editor
cd spark-workshop-template
code .

# 3. Customize CLAUDE.md with your org's name and instructions
# 4. Fill in state/org-state.md with your real projects and priorities
# 5. Design your agents in agents/ (start with 3, expand to 5-8)
# 6. Add 3-5 initial memories to memory/

# 7. Start Claude Code and load your project
claude

# 8. Your first prompt:
# "Read CLAUDE.md and the state files. What's the status of my projects?"

Your First Session

Ask Spark:

• "What's the status of my projects?"
• "Draft a grant strategy for [funder]"
• "Review this [document] for [criteria]"
• "Draft an email to [person] about [topic]"
• "What decisions have we made this month?"
• "What's stale? What needs attention?"

The Architecture

┌─────────────────────────────────────────┐
│                 YOU                      │
│                    │                     │
│                    ▼                     │
│               ┌─────────┐               │
│               │  SPARK   │  ← Router    │
│               │  (CAO)   │               │
│               └────┬─────┘               │
│         ┌──────────┼──────────┐          │
│         ▼          ▼          ▼          │
│    ┌─────────┐ ┌─────────┐ ┌─────────┐  │
│    │ Agent 1 │ │ Agent 2 │ │ Agent 3 │  │
│    └────┬────┘ └────┬────┘ └────┬────┘  │
│         │          │          │          │
│         ▼          ▼          ▼          │
│    ┌────────────────────────────────┐    │
│    │     STATE FILES (markdown)     │    │
│    │  portfolio · priorities · log  │    │
│    │  decisions · people · memory   │    │
│    └────────────────────────────────┘    │
│                    │                     │
│              ┌─────┴─────┐               │
│              │  TOOLS    │               │
│              │ Gmail MCP │               │
│              │ Calendar  │               │
│              │ Git/Deploy│               │
│              └───────────┘               │
└─────────────────────────────────────────┘

Key Principles

1. You learn by doing it, because you are capable of doing it. Don't read about Spark. Build your Spark. The template is a trail, not a textbook. Walk it.
2. 8 agents max. Google research shows multi-agent systems lose performance above 8. Make each one smarter, not more numerous.
3. State files ARE the institution. If the AI forgets everything between sessions, it's a stranger you brief every morning.
4. Drafts only, never sends. The AI drafts emails. The human reviews and sends. Tone is personal. Relationships are human.
5. The protocol is the discipline. Every session starts the same (read state, check contradictions) and ends the same (update state, draft morning email).
6. Memory is earned. The AI remembers corrections, preferences, and context across sessions. It gets better the more you use it.

Workshop

Want to build your Spark with guided help?

• 60-min Keynote: The story of how Spark was built, the architecture, live demos, Q&A
• 3-hour Workshop: Hands-on build. You leave with a working system.
• Full-day Intensive: Advanced features — cross-project intelligence, contradiction detection, scenario modeling

Workshop materials and facilitator guides: radicalimagination.xyz/spark (password-protected for workshop attendees)

Book a workshop: create@radicalimagination.xyz

Built By

Radical Imagination — a collective that designs VR/AR/XR experiences honoring cultural heritage. Spark is the operating system. Now it's yours.

---

"The system didn't make the art. The system made the art possible."