P-5: Handle Dialogue

[spuentesp]

Category: play | Epic: 4 | Priority: high

Summary

Process player dialogue with NPCs or environment. Loads NPC personality,
memories, and facts to generate contextually appropriate responses.
Creates memories for both PC and NPC.

Acceptance Criteria

• NPC identified from dialogue target
• NPC personality loaded for response generation
• Semantic memory search retrieves relevant NPC memories
• NPC response reflects personality, knowledge, relationships
• Memory created for NPC about this conversation
• Memory created for PC if significant information exchanged
• Relationship changes proposed if applicable
• Information revelation tracked

Dependencies

This use case depends on:

• DL-2
• DL-7
• DL-10
• P-3

Implementation

Layer: 2

Files to modify:

• packages/agents/src/monitor_agents/narrator.py
• packages/agents/src/monitor_agents/memory_manager.py

NEO4J Operations:

• neo4j_get_entity (authority: *)
• neo4j_list_facts (authority: *)

MONGODB Operations:

• mongodb_get_memories (authority: *)
• mongodb_create_memory (authority: *)

QDRANT Operations:

• qdrant_search_memories (authority: *)

Notes:

• NPC consistency is key
• Consider mood modifiers
• Track what information NPCs reveal

Testing Requirements

Minimum coverage: 80%

Unit tests:

• test_dialogue_npc_identified: target parsed
• test_npc_memories_retrieved: relevant memories found
• test_npc_response_personality: response matches personality
• test_memory_created: NPC memory stored
• test_dialogue_narration: non-NPC speech handled

Integration tests:

• test_dialogue_full_flow: dialogue -> response -> memory
• test_npc_information_tracking: info exchange tracked

References

Documentation:

• docs/USE_CASES.md#p-5-handle-dialogue
  Code:
• packages/agents/src/monitor_agents/narrator.py
• packages/agents/src/monitor_agents/memory_manager.py

---

Generated from /home/sebas/monitor2/docs/use-cases/play/P-5.yml