Generative Layers
A Java framework for governed generative resource layers around agent-oriented systems.
Purpose
Generative Layers introduces a governed resource-layer abstraction for BDI agent systems.
It situates generative resources within an external body layer around the agent, expanding the design space for governed generative use without altering the underlying reasoning model.
Target platforms
ASTRA
AgentSpeak(TR)-style BDI target.
Jason
AgentSpeak-style BDI target.
JaCaMo
Jason, CArtAgO, and Moise integration.
Integration Details & Specifications
Governance architecture, target platforms, and custom integration adapters.
Generative Layers is built to support governed integration with prominent BDI (Belief-Desire-Intention) agent languages and multi-agent programming platforms. The core framework currently offers implemented cross-platform integrations for ASTRA, Jason, and the integrated JaCaMo platform (which binds Jason BDI agents, CArtAgO environments, and Moise organizations together).
Compare how the same governance flow (configuring, invoking, validating, and adopting candidate material) is realized natively across each target agent environment:
ASTRA Module
Implemented as a native ASTRA Module written in Java. Commands are mapped to inline custom statements.
Jason Internal Actions
Implemented as standard AgentSpeak Internal Actions in Java. Invoked natively by BDI agents using custom namespace extensions.
CArtAgO Artifact (JaCaMo)
Implemented as a CArtAgO Environment Artifact in Java. Commands are operations called once focused on the artifact, receiving updates natively via observable belief signals.
To view complete syntax comparisons, code snippets, and learn how integrations are realized natively under the hood, explore the Platform syntax comparison developer reference guide.
Interested in connecting Generative Layers to another BDI agent language, Multi-Agent System (MAS) framework, or custom agent platform? We warmly welcome new integrations and adapters! For public technical discussion, bug reports, feature requests, or adapter proposals, please open an issue in the Generative Layers framework repository. You can also contact us at [email protected].