run-scenario

Name: run-scenario
Availability: InStock
Author: swarm-ai-safety

Execute SWARM simulation scenarios and export history.json and CSV metrics.

39|4|Updated Feb 3, 2026

Search Tags:#benchmarking #multi-agent #simulation #swarm #scenario #risk analysis

Ready to run

npx skills add https://github.com/swarm-ai-safety/swarm --skill run-scenario

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: run-scenario
Source: https://github.com/swarm-ai-safety/swarm/tree/main/bench/skills/run-scenario
Command: npx skills add https://github.com/swarm-ai-safety/swarm --skill run-scenario

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill automates the execution of SWARM simulation scenarios, ensuring reproducible results and standardized artifact generation for risk analysis.

Core Features & Use Cases

Scenario Execution: Runs SWARM simulations with specified seeds, epochs, and steps.
Artifact Export: Generates history.json and CSV files for detailed analysis.
Use Case: A researcher needs to baseline the performance of a new multi-agent setup. They use this Skill to run the baseline scenario with a specific seed and collect all the output metrics for comparison.

Quick Start

Use the run scenario skill to execute the baseline scenario with seed 42 and save the artifacts to the /tmp/output directory.

Frequently Asked Questions about run-scenario

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I run multi-agent SWARM simulations and export the results for risk analysis?▼

To run multi-agent SWARM simulations, execute scenario scripts with specific seeds and epochs to generate standardized risk analysis artifacts. The execution exports a full history.json file and a dedicated CSV directory containing detailed output metrics for reproducible benchmarking.

What is the best way to ensure reproducible benchmarking results across multiple simulation scenarios?▼

Reproducible benchmarking results require executing SWARM simulation scenarios with a specified seed, epoch, and step count. This standardized scenario execution generates consistent history.json and CSV metrics, ensuring that multi-agent risk analysis outputs remain identical across repeated programmatic or CLI runs.

Can I execute SWARM scenario simulations programmatically, or do I need to use the command line interface?▼

SWARM scenario simulations support both programmatic execution and command line interface usage. Scenario path resolution handles the input configuration, while artifact organization automatically directs the exported history.json and CSV metrics into a specified output directory regardless of the execution method chosen.

How do I collect and organize CSV metrics from a baseline multi-agent simulation scenario?▼

Collecting CSV metrics from a baseline multi-agent simulation involves running the scenario and exporting artifacts to a dedicated CSV directory. The standardized export process automatically organizes the detailed output metrics alongside the full simulation history.json file for immediate comparative analysis.

Does the SWARM scenario execution tool require any external dependencies to generate history and metrics artifacts?▼

SWARM scenario execution requires no external dependencies to generate history.json and CSV metrics artifacts. The standalone scripts handle scenario path resolution, simulation execution, and artifact organization natively, allowing multi-agent risk analysis benchmarking to run in an isolated environment.