ArchitectureA supervised JVM-class runtime — OLTP on seven engines, OLAP on three. AI-native, MCP-native, observable as plain SQL.Read the architecture
Está viendo la edición Perú. Está viendo la edición Colombia. You're viewing the Pakistan edition. Cambiar a la edición global →Cambiar a la edición global →Switch to the global edition →
AI and agentsProduct Release1 min read

Agents and skills defined in the dictionary, skills offered as slash commands

An agent is instantiated from a dictionary record naming its prompt, tools, skills and iteration limit, and the orchestrator registers it as a sub-agent at runtime. Skills are fetched on demand.

An agent that lives in code ships with the platform. An agent that lives in the dictionary ships with the application, under the application's own versioning and grants. This release makes the second kind possible.

Agents as records

  • Instantiated from the dictionary An agent record carries its prompt, its tools, its skills and its iteration limit. The orchestrator registers each as a sub-agent dynamically.
  • Tools and skills as shared resources, cached by code, so several agents reference one definition.
  • Scriptable A dictionary agent can be created and run from server-side JavaScript.

Skills

  • Catalogued to the model The skills catalogue is appended to the system prompt as codes and descriptions; the full text arrives only when the model asks for one.
  • Slash commands The chat exposes the active database's skills as slash commands with their descriptions.

An application team can now ship an agent the way it ships a report: as a definition, reviewed and journalled, without a platform release.

On the platformAI & MCPPlatform

All posts