[FEATURE] Story teller system for simulation content generation

[ncudlenco]

Is your feature request related to a problem? Please describe.
Currently, there is no automated system to generate natural language stories based on the simulation's assets, such as skins, objects, contexts (locations), and possible actions. This limits the ability to create rich, narrative-driven datasets and content for downstream tasks (e.g., language modeling, scenario generation, or annotation).

Describe the solution you'd like
Create a storyteller system that automatically writes down (generates) stories based on the available skins, objects, contexts (locations), and actions present in the simulation. The system should:

• Analyze all available assets (skins, objects, contexts, actions)
• Generate narrative stories that combine these elements into plausible sequences or scenarios
• Provide output in a structured format, e.g., plain text or JSON, for downstream consumption
• Optionally, allow for parameterization (e.g., story length, focus object, location)

Describe alternatives you've considered

• Manual story creation (not scalable or reproducible)
• Using external story generators with no knowledge of in-game context

Acceptance Criteria

• Story teller system generates stories based on available skins
• Story teller system generates stories based on available objects
• Story teller system generates stories based on available contexts (locations)
• Story teller system generates stories based on available actions
• Generated stories are plausible and combine elements from the simulation
• Output format is documented and suitable for downstream use
• Documentation and usage instructions are provided

Additional context