fleet
por oguzhnatly
Coding Agents & IDEsResumo
Monitor, manage, and orchestrate a fleet of OpenClaw agent gateways.
Descrição
O que esta skill faz
CLI toolkit for coordinating a fleet of OpenClaw agent gateways. A coordinator agent uses it to dispatch tasks, monitor live sessions, steer agents mid-task, and compute reliability scores from dispatch history. All operations run locally over loopback, with optional GitHub CI and Linear ticket integration.
Without fleet, managing multiple agent gateways requires manually polling each one, maintaining dispatch logs separately, and computing reliability from raw session files.
Quando usá-la
- Dispatch a coding task to a specific agent and stream its output live
- Run a morning fleet report to catch overnight CI failures and agent outages
- Send a mid-task correction to an agent currently working on a feature
- Check trust scores before assigning a critical deployment to an agent
- Decompose a large task and run subtasks across multiple agents in parallel
Exemplo de fluxo de trabalho
User asks: check the fleet, then dispatch a pagination task to the coder agent and monitor it
- 1
Run fleet health to verify all gateways and configured endpoints are online
- 2
Run fleet trust to review coder's current reliability score
- 3
Dispatch the task with fleet task coder "add pagination to /api/spots"
- 4
Monitor live output with fleet watch coder
Health check passes, trust score confirmed, task dispatched and streamed in real time to completion
Requisitos
- LINEAR_API_KEY environment variable (for Linear ticket integration)
- GitHub account with authenticated gh CLI (for fleet ci command)
Instalação
clawhub install oguzhnatly/fleetExecute o comando da CLI ou cole o prompt no Claude Code, Cursor, Codex ou no agente de código que você usa.