Orchestrate generative agent simulations with host-resident architecture and blackboard communication.
npx skills add https://github.com/zc6600/aura --skill stanford-town
Please help me install this Agent Skill. Skill: stanford-town Source: https://github.com/zc6600/aura/tree/main/use-cases/smallville-town/skills/stanford-town Command: npx skills add https://github.com/zc6600/aura --skill stanford-town
This skill solves the challenge of simulating complex, emergent social behaviors in multi-agent systems without relying on hardcoded scripts or prompt leakage.
Initialize the stanford-town simulation environment to begin the hourly agent dispatch loop.
High-intent search queries and answers about installing and using this skill.
You can simulate emergent social behavior by decoupling environment state from individual agent decision-making, using a host-resident architecture with blackboard and mailbox communication patterns to facilitate complex social emergence autonomously.
Autonomous generative agents make decisions based on personal memory streams and mailbox notifications, operating as independent subagents that interact and form relationships without relying on narrative prompt leakage.
Orchestrating multi-agent simulations with isolated state requires subagent, blackboard, mailbox, and groupchat modules to maintain separation between global environment state and individual agent decision-making.
Yes, the simulation uses a Town Host architecture that manages global environment state and time synchronization via a blackboard system, ensuring agents react to a consistent shared reality.
To start a generative agent simulation, initialize the simulation environment to begin the hourly agent dispatch loop, allowing spawned resident subagents to perform daily routines autonomously based on their personas.
Prompt leakage in multi-agent orchestration is prevented by maintaining isolated state through a host-resident architecture, ensuring resident subagents rely on memory streams rather than exposed narrative prompts.
MCERQUAGuide desktop agents through nightly reflection and post summaries to the Blackboard.
ProntingAnalyze chat logs to extract emotional states, intentions, and power dynamics.
GeargrindadminOrchestrate multi-agent swarms with sequential, concurrent, and hierarchical patterns.
brkastnerDelegate ClickUp task API calls to the clickup-task-agent via /task commands.
