Datacore Org

A framework for autonomous organization management.

Vision

Datacore Org is the foundation for companies that run themselves. Agents handle day-to-day operations while humans focus on:

• Teaching - Training agents on domain expertise and preferences
• Supervising - Reviewing outputs, correcting course
• Deciding - Making strategic choices that require human judgment

How It Works

Human Input                    Agent Operations
───────────                    ────────────────

Strategy & Vision    ───────>  Planning & Execution
                              ├── Product roadmaps
                              ├── Content generation
                              ├── Research & analysis
                              └── Coordination

Teaching & Feedback  ───────>  Learning & Improvement
                              ├── Pattern recognition
                              ├── Style adaptation
                              └── Process optimization

Decisions            <───────  Proposals & Options
                              ├── "Here are 3 approaches..."
                              ├── "This needs your approval..."
                              └── "Escalating: blocker found..."

Departmental Modules

Each department can have its own module with specialized agents:

Module	Purpose	Status
datacore-comms	Marketing, content, social media	Planned
datacore-dev	Engineering workflows, deployments	Planned
datacore-product	Product management, specs, roadmaps	Planned
datacore-research	Market research, competitor analysis	Planned
datacore-ops	Operations, OKRs, coordination	Planned
datacore-finance	Financial tracking, reporting	Planned

These modules will be developed primarily by teams using them (e.g., Datafund team developing their areas).

Quick Start

# 1. Fork this template to your org
# github.com/datacore-one/datacore-org → your-org/space

# 2. Clone to your Datacore installation
cd ~/Data
git clone [email protected]:your-org/space.git 1-your-org

# 3. Customize CLAUDE.md with your org context

# 4. Install departmental modules as needed

Structure

[N]-[name]/                       # e.g., 1-datafund/
├── CLAUDE.md                     # Org context for AI
├── _index.md                     # Space home page
│
├── .datacore/
│   ├── config.yaml               # Space configuration
│   ├── commands/                 # Org-specific commands
│   ├── agents/                   # Org-specific agents
│   ├── learning/                 # AI learning (patterns, preferences)
│   └── modules/                  # Departmental modules (gitignored)
│
├── org/                          # Task management (GTD)
│   ├── inbox.org                 # Team task capture
│   └── next_actions.org          # AI task queue
│
├── 0-inbox/                      # Unprocessed notes
│
├── journal/                      # Team activity log
├── today/                        # Daily briefings
│
├── 1-departments/                # Active work by department
│   ├── ops/                      # Operations
│   ├── product/                  # Product management
│   ├── dev/                      # Engineering
│   ├── research/                 # Research
│   └── comms/                    # Communications
│
├── 1-projects/                   # Code repos (gitignored)
│
├── 2-knowledge/                  # Shared knowledge (Zettelkasten)
│   ├── insights.md               # Organizational insights
│   ├── zettel/                   # Atomic concepts
│   ├── literature/               # Source summaries
│   ├── reference/                # People, companies, glossary
│   └── [topic]-docs/             # Topic documentation
│
└── 3-archive/                    # Historical content

Numbered Folder Convention

Numbers indicate processing stage:

• 0- = Capture/inbox
• 1- = Active work (departments, projects)
• 2- = Permanent knowledge
• 3- = Archive

Index Files

Every major folder has _index.md providing:

• Navigation and entry points
• Status summaries
• Links to key content

Agent Roles

Core agents (built into datacore-org):

• org-coordinator - Routes tasks, manages handoffs between departments
• daily-briefing - Generates /today briefings
• standup-generator - Compiles team updates

Departmental agents (from modules):

• Each module brings specialized agents for its domain
• e.g., datacore-comms brings content writers, social schedulers
• e.g., datacore-dev brings code reviewers, deployment managers

AI Learning

Agents improve over time through .datacore/learning/:

File	Purpose
patterns.md	Successful approaches to remember
corrections.md	Human feedback and corrections
preferences.md	Org style, voice, preferences

Human Touchpoints

Agents escalate to humans for:

• Strategic decisions
• External communications (final approval)
• Budget/resource allocation
• Hiring decisions
• Legal/compliance matters
• Anything flagged as "needs human judgment"

The Autonomous Company Roadmap

1. Phase 1: Assisted (Current)

   • Agents draft, humans review and publish
   • Agents research, humans decide

2. Phase 2: Supervised

   • Agents execute routine tasks autonomously
   • Humans review summaries, handle exceptions

3. Phase 3: Autonomous

   • Agents run day-to-day operations
   • Humans set direction and handle edge cases

4. Phase 4: Self-Improving

   • Agents identify process improvements
   • Agents train other agents
   • Humans focus purely on strategy

Contributing

Departmental modules are developed by teams using them. If you're building out a department:

1. Create module in datacore-one/datacore-[dept]
2. Add agents for department-specific tasks
3. Add commands for department workflows
4. Document in module README

License

MIT