simulation

name	simulation
description	Abstract base for runtime state — adventure, city-sim, ecosystem all inherit from this
license	MIT
tier	1
allowed-tools	["read_file","write_file"]
related	["adventure","society-of-mind","time","party","character","needs","buff","action-queue","prototype","speed-of-light","exit","memory-palace","simulator-effect"]
tags	["moollm","runtime","state","turns","engine"]

Simulation Skill

Abstract base for runtime state management. Concrete simulations (adventure, city-sim, ecosystem) inherit from this.

Motto: "The simulation is the world. The world is the simulation."

[!IMPORTANT] This is the abstract base. Don't create SIMULATION.yml directly. Use ADVENTURE.yml (for narrative exploration) or create your own concrete type. The concrete file INCLUDES all simulation properties plus type-specific ones.

Key Concepts

• SIMULATION.yml — Source of truth for "now"
• Global parameters — Configurable via chat
• Git time machine — Commits = deterministic undo
• Turn tracking — Increments on significant actions

Global Parameters

Time Control

Parameter	Default	Chat Command
paused	false	PAUSE / RESUME
advancement	normal	SET TIME ADVANCEMENT [none|slow|normal|fast]
turn	0	TICK [n] / REWIND [n]

Git Automation

Parameter	Default	Chat Command
auto_commit	false	SET AUTO COMMIT [on|off]
auto_push	false	SET AUTO PUSH [on|off]
git_remote	origin	SET GIT REMOTE [name]
git_branch	main	SET GIT BRANCH [name]

Display

Parameter	Default	Chat Command
narration_level	normal	SET NARRATION [minimal|normal|verbose]
show_mechanics	false	SET SHOW MECHANICS [on|off]
debug_mode	false	DEBUG [on|off]

Output (all ON by default)

Parameter	Default	Chat Command
write_to_chat	true	SET CHAT OUTPUT [on|off]
write_to_transcript	true	SET TRANSCRIPT [on|off]
transcript_path	./README.md	SET TRANSCRIPT PATH [path]
include_yaml_islands	true	SET YAML ISLANDS [on|off]
include_links	true	SET LINKS [on|off]
link_format	markdown	SET LINK FORMAT [markdown|plain|none]

Transcript paths are relative to adventure directory:

• ./README.md (default)
• ./logs/session-log.md
• ./narrative/chapter-1.md

Example output with all flags ON:

You're in the pub with Marieke.

You examine the lamp:

object:
  id: brass-lamp
  state: lit
  oil: 75%
  # Warm. Faithful. Flickering slightly.

Gameplay

Parameter	Default	Chat Command
difficulty	normal	SET DIFFICULTY [easy|normal|hard]

SIMULATION.yml Structure

simulation:
  turn: 47
  paused: false
  
parameters:
  git:
    auto_commit: true
    auto_push: false
    
  gameplay:
    difficulty: normal
    permadeath: false
    
  display:
    show_mechanics: false
    
player:
  character: characters/don-hopkins/
  location: pub/
  
party:
  members: [characters/don-hopkins/, pub/cat-cave/terpie.yml]
  
selection:
  targets: []
  
active_buffs: []
world_state: {}
flags: {}

Commands

Command	Effect
STATUS	Current simulation state
SHOW SETTINGS	List all parameters
SET [param] [value]	Configure parameter
GET [param]	Query parameter value
PAUSE / RESUME	Time control
TICK [n]	Force advance turns
SAVE / LOAD	Persistence

Git Time Machine

With auto_commit: true, every turn is a git commit:

# Commits look like:
Turn 47: Entered pub
Turn 48: Ordered Monkey's Blessing
Turn 49: Pet Terpie (+Serenity buff)

Time travel commands:

• REWIND [n] — Go back n turns
• BRANCH [name] — Create alternate timeline
• TIMELINES — List all branches
• MERGE TIMELINE [branch] — Merge narratives

See Also

• ONE-STEP-TRAP.md — Sutton temporal abstraction vs micro-rollout; Urban Safari field case
• examples/one-step-trap.yml — machine-readable skill cross-links
• time — Turn tracking, duration
• party — Party and selection
• buff — Active effects
• character — Player state
• adversarial-committee — Decision-making within simulation
• speed-of-light — Multi-agent simulation in one call