Openclaw Skill Gastown

What This Skill Does

Gas Town is a sophisticated multi-agent orchestration framework designed specifically for Claude Code, serving as a 'Cognition Engine' that transforms disjointed AI interactions into a persistent, structured workflow. Unlike standard agent setups where context evaporates upon session termination, Gas Town introduces a state-managed environment where work items are tracked as 'Beads' within a git-backed ledger. It acts as an operational layer, allowing developers to manage teams of 20-30 agents by providing persistent identities, internal communication channels, and automated monitoring via the Witness, Refinery, and Mayor subsystems. The skill automates complex task distribution, ensuring that every piece of work is accounted for, routed to the correct agent, and monitored through a reliable hook-based architecture.

Installation

To integrate the Gas Town engine into your environment, run the following command within your OpenClaw terminal: clawhub install openclaw/skills/skills/saesak/openclaw-skill-gastown Ensure that you have your workspace initialized as a git repository, as Gas Town relies on git-backed hooks to maintain its persistence layer.

Use Cases

Gas Town is ideal for complex, long-running software engineering projects. It excels in scenarios requiring architectural migration, large-scale refactoring, or multi-module feature development where context-switching between agents leads to productivity loss. It is specifically built for users who need to hand off tasks between specialized agents without losing the thread of progress or the underlying rationale behind code changes.

Example Prompts

1. "Initialize a new project environment, set up the agent rigs, and create a tracking bead for the upcoming authentication service migration."
2. "Review the current status of all active convoys; I need to see which agents have completed their assigned tasks and which are awaiting oversight from the Witness."
3. "Sling the 'Database Optimization' bead to the primary dev-rig and ensure the patrol is activated so the Mayor can monitor the output."

Tips & Limitations

• Proactive Monitoring: Regularly use gt status and gt doctor to ensure the communication pipeline remains synchronized. Do not wait for an error to check the health of your agent rigs.
• Metaphorical Precision: The system relies on its internal identity framework. Embracing the 'Gas Town' metaphors (Beads, Convoys, Rigs) helps the AI maintain better context for system state.
• Boundaries: Remember that Gas Town manages the orchestration, but you remain the commander. You must trigger task handoffs and verify agent outputs. The system is designed to scale your capability, not replace your judgment.