gastown

Coordinate multi-agent workspaces with Gas Town commands and Dolt server operations.

Updated Jan 2, 2026
One-click install
npx skills add https://github.com/r0man/gastown.el --skill gastown-r0man
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gastown
Source: https://github.com/r0man/gastown.el/tree/main/.claude/skills/gastown
Command: npx skills add https://github.com/r0man/gastown.el --skill gastown-r0man

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Gas Town provides a centralized orchestration layer to coordinate multi-agent workspaces, enabling reliable spawning, communication, and coordination of rigs, polecats, witnesses, refiners, and related components.

Core Features & Use Cases

  • Orchestrates agent roles (Polecat, Witness, Refinery, Mayor, Deacon, Overseer) across rigs
  • Dispatches work and manages context with gt sling, hook, and related commands
  • Enables ephemeral and persistent communication via gt nudge and gt mail, plus Dolt server operations
  • Manages sessions, health, and checkpointing to recover work after restarts
  • Provides a reference architecture for rig layout and component interactions

Quick Start

Load Gas Town context, inspect the rig, and begin deploying initial polecat sessions with gt sling to start a project.

Frequently Asked Questions about gastown

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

FAQPage Schema
How do I orchestrate multi-agent workspaces for coordinated task execution?

You orchestrate multi-agent workspaces by deploying agent roles like polecats and witnesses across rigs, dispatching work with commands like gt sling, and enabling communication via gt nudge and gt mail for coordinated task execution.

What is the best way to dispatch work and manage context across multiple agents?

The best way to dispatch work and manage context is using the gt sling command, which distributes tasks to spawned agent sessions while maintaining session health and checkpointing to recover work after restarts.

How do agents communicate with each other in a multi-agent rig setup?

Agents in a multi-agent rig setup communicate through ephemeral and persistent channels, specifically using gt nudge for immediate signals and gt mail for longer-lasting messages, plus Dolt server operations for shared state.

Can I recover agent work and sessions after unexpected restarts?

Yes, you can recover agent work after unexpected restarts because the orchestration layer manages sessions, health, and checkpointing, allowing rigs, polecats, and refiners to resume operations from their last saved state.

Do I need YAML frontmatter to activate multi-agent orchestration workflows?

Yes, you need YAML frontmatter containing name and description fields to activate multi-agent orchestration workflows, and you can optionally include scripts, references, or assets to extend the agents' behavior.

What roles can I assign to agents when coordinating a multi-agent workspace?

You can assign specialized roles such as Polecat, Witness, Refinery, Mayor, Deacon, and Overseer when coordinating a multi-agent workspace, each interacting differently across rigs to process and refine dispatched work.