bdi-mental-states

The bdi-mental-states skill provides a framework for integrating formal cognitive architectures into AI agents. By utilizing a BDI (Belief-Desire-Intention) model, this skill allows agents to maintain a persistent, explainable mental state, transforming raw RDF context data into actionable cognition. It is designed for developers building rational agency, multi-agent systems, or neuro-symbolic AI applications where traceability and logical consistency are critical. This skill bridges the gap between semantic web data structures and LLM reasoning cycles, enabling sophisticated deliberative behaviors that standard prompt-based interactions often lack.

• Formal ontology pattern implementation for defining Beliefs, Desires, and Intentions as endurants.

• Modeling of mental processes (BeliefProcess, DesireProcess, IntentionProcess) as perdurants for causal reasoning and auditability.

• Support for T2B2T (Triples-to-Beliefs-to-Triples) pipelines to consume RDF data and produce semantic output.

• Goal-directed planning integration using 'specifies' and 'precedes' properties for task decomposition.

• Justification attachment (bdi:isJustifiedBy) to provide explainable reasoning chains for every mental entity.

• C4-level notation guidance to align mental models with structural system architecture.

• Ideal for use cases involving JADE, JADEX, or SEMAS frameworks and general cognitive architecture research.

• Ensure all beliefs are grounded in specific WorldState references to maintain semantic interoperability.

• Use the provided bidirectional properties (motivates/fulfils) to enable both forward-chaining and backward-tracing reasoning.

• Suitable for Logic Augmented Generation (LAG) tasks where neuro-symbolic integration improves reliability.

• Constrain mental states by time intervals using formal ontology patterns to prevent state ambiguity.

• Always define the Plan execution lifecycle clearly to allow agents to transition from deliberation to action seamlessly.