The Domain Layer

Expertise, packaged and deployable.

A skill encodes what an expert knows how to do — the research sequence, the quality checks, the output format, the domain-specific judgment calls — into a form that AI agents can execute reliably. An AI workspace bundles skills with data and tools into a turnkey environment.

Anatomy of a Skill

Four components. One unit of expertise.

Each skill is self-contained — it carries its own instructions, data context, tool permissions, and output templates. No external dependencies, no configuration drift, no "works on my machine" problems.

Instructions

Step-by-step workflow guidance — decision trees, quality gates, and output specifications that encode domain expertise.

Data Context

Company-specific data — brand identity, stakeholder profiles, market context — loaded automatically when the skill activates.

Tool Access

Curated set of MCP tools and capabilities the skill is authorised to use, with safety boundaries and permission gates.

Output Templates

Format specifications for deliverables — document structures, chart types, data schemas — that ensure consistent, professional output.

AI Workspaces

Skills don't ship alone.

An AI workspace is the complete package — skills, data connectors, brand assets, and instructions deployed as a single installable unit. Your team gets a fully configured AI environment.

Turnkey Deployment

Install a workspace with a single command. All skills, data connectors, brand assets, and instructions arrive as a self-contained package. No configuration required — it just works.

Marketplace Distribution

Workspaces are distributed via marketplace — org admins install once, every team member gets access. New skills and updates propagate automatically.

Permission Boundaries

Each skill declares exactly which tools it needs and what actions it can take. Write operations require explicit human approval. No skill can exceed its declared scope.

Zero-Code Operation

Partners use skills through natural language. The AI agent handles tool selection, data retrieval, analysis sequencing, and output formatting — partners focus on strategic direction.

Deployment Models