How does Gas Town handle stuck agents?

The system utilizes the 'Witness' role, which monitors worker progress and automatically intervenes if an agent becomes blocked or unresponsive.

What is the 'Refinery' in this system?

The Refinery is a supervisory role responsible for quality control and merging completed work from agents back into the main codebase.

What is a 'Polecat' in Gas Town?

A Polecat is a specialized, short-lived AI worker spawned to handle a specific task and then vanish once the work is completed or merged.

Can I run multiple tasks at once?

Yes, Gas Town is built for parallel execution. You can 'sling' multiple task beads to different rigs simultaneously to increase development throughput.

Do I need to run terminal commands manually?

No. Gas Town is designed so that the Overseer (you) provides natural language instructions, and the skill executes all necessary gt and bd CLI commands via the Bash tool.

Gas Town Multi-Agent Orchestrator

Name: Gas Town Multi-Agent Orchestrator
Author: numman-ali

bynumman-ali

•

825

•

생산성 및 워크플로

Orchestrates multiple AI agents to execute complex, parallel workflows through an automated 'sling and hook' engine.

Gas Town is a high-performance orchestration engine designed to scale Claude Code's capabilities by managing a fleet of specialized AI agents. It transforms your development environment into an automated 'software factory' where you act as the Overseer. The system handles the entire lifecycle of multi-agent work—from creating 'beads' (tasks) and 'slinging' them to workers like 'polecats' or 'crew,' to automated monitoring by the Witness and merging by the Refinery. This skill eliminates manual terminal management and context switching, allowing developers to drive high-throughput parallel pipelines with automated crash recovery and supervisor oversight.

주요 기능

01Automated task tracking using beads, hooks, and convoys

02Automated supervisor roles (Witness, Refinery, Mayor) for monitoring and merging

03825 GitHub stars

04Multi-agent orchestration and worker lifecycle management

05Seamless execution of gt and bd CLI operations without manual terminal entry

06Parallel task execution with specialized 'polecat' and 'crew' agents

사용 사례

01Managing large-scale repository refactors using specialized agent roles

02Automating complex workflows involving task tracking, monitoring, and quality control

03Executing multiple bug fixes or features in parallel across different project rigs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add numman-ali/n-skills gastown

For use in Claude.ai and ChatGPT