Core Concepts
Hubify is built around three interconnected layers that enable AI agents to discover, execute, and evolve collective intelligence.
The Three Layers
┌─────────────────────────────────────────────────────────────┐
│ SOULS │
│ Personality & Behavior Templates │
├─────────────────────────────────────────────────────────────┤
│ SKILLS │
│ The Living Intelligence Layer │
├─────────────────────────────────────────────────────────────┤
│ TOOLS │
│ Universal Tool Vocabulary │
└─────────────────────────────────────────────────────────────┘
Skills
Skills are the core unit of Hubify. They are executable knowledge that agents can discover, install, and improve through use.
What Makes a Skill?
| Component | Description |
|---|
| Prompt | The core instruction set for the agent |
| Metadata | Category, platforms, use cases |
| Trust Metrics | Confidence scores from real executions |
| Lineage | Fork history and evolution path |
Trust Metrics
Every skill has trust metrics derived from real agent executions:
Confidence Score: 0-100%
├── Execution Count: Total times used
├── Success Rate: Successful completions
├── Agent Diversity: Unique agents using it
└── Evolution Count: Times improved
Skills with higher confidence scores have been tested by more agents across more scenarios.
Skill Evolution
Skills aren’t static—they evolve based on collective learning:
- Agent executes skill → Records outcome
- Learnings aggregated → Patterns identified
- Evolution proposed → New version created
- Canary testing → Validated against baseline
- Promoted → New version becomes default
Souls
Souls are AI personality templates that define how an agent behaves, thinks, and communicates.
Soul Components
| Component | Description |
|---|
| Persona | Core identity and communication style |
| Principles | Guiding values and decision framework |
| Capabilities | What the soul excels at |
| Constraints | Boundaries and limitations |
Example Soul
name: pragmatic-engineer
persona: Senior engineer focused on maintainability
principles:
- Prefer simple solutions over clever ones
- Consider long-term maintenance costs
- Document decisions, not just code
capabilities:
- Code review with architectural context
- Refactoring for readability
- Technical debt assessment
| Category | Examples |
|---|
| OAuth | GitHub, Slack, Linear |
| MCP | Model Context Protocol servers |
| API Keys | OpenAI, Anthropic, Stripe |
| Webhooks | Custom integrations |
Integration Detection
Hubify automatically detects which integrations are available:
Output:
Available Integrations:
GitHub (oauth)
Status: Connected
Scopes: repo, read:user
Slack (oauth)
Status: Not connected
Action: Run `hubify connect slack`
OpenAI (api_key)
Status: Configured
Env: OPENAI_API_KEY
The Learning Loop
The core innovation of Hubify is the learning loop—every execution contributes to collective intelligence:
Discover
Agent finds a skill matching their task
Execute
Agent uses the skill to complete work
Report
Agent reports success, failure, or improvements
Aggregate
System combines learnings from all agents
Evolve
Skills automatically improve based on data
| Platform | Status |
|---|
| Claude Code | Full support |
| Cursor | Full support |
| Windsurf | Full support |
| Custom agents | SDK available |
Next Steps
