bdi-mental-states

This skill provides a formal framework for modeling agent mental states, specifically designed for developers building cognitive architectures or rational agents. By implementing the Belief-Desire-Intention (BDI) paradigm, it allows LLM-based agents to transition from reactive prompt-response cycles to deliberative reasoning processes that are traceable and explainable. It is intended for researchers, engineers, and developers working on multi-agent systems, neuro-symbolic AI, and semantic web integrations where the agent must maintain long-term internal states and follow explicit goal-directed logic.

• Transform raw external RDF context into structured cognitive entities (Beliefs, Desires, Intentions) using formal BDI ontology patterns.

• Enable bidirectional 'Triples-to-Beliefs-to-Triples' (T2B2T) pipelines to ensure semantic consistency between external world states and internal mental states.

• Support complex agent reasoning patterns including perception-action cycles, goal decomposition, and temporal evolution of mental states.

• Facilitate explainability by linking every mental state to its grounding world state reference and its generating causal processes via the bdi:Justification mechanism.

• Compatible with formal BDI frameworks like SEMAS, JADE, and JADEX for heterogeneous multi-agent collaboration.

• Activate this skill when the agent needs to perform multi-step planning, maintain belief stores, or operate within frameworks requiring logic-augmented generation.

• Input expected is usually RDF or graph-based context data; output is typically actionable agent plans or updated belief instances.

• Use the provided bidirectional properties like motivates, fulfils, and specifies to maintain the structural integrity of the mental chains.

• Always ground mental states in specific WorldState references to prevent ungrounded belief formation that could lead to hallucination or logical inconsistency.

• Constraint: Agents must be capable of processing structured schema data or have access to RDF-parsing capabilities within their reasoning context to fully utilize the ontology patterns.