agent-fleet

Monitor multi-agent fleet lifecycles with budgets, circuit breakers, and heartbeat detection.

6|Updated Feb 26, 2026
One-click install
npx skills add https://github.com/Acosmi/CrabClaw --skill agent-fleet
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-fleet
Source: https://github.com/Acosmi/CrabClaw/tree/main/docs/skills/operations/agent-fleet
Command: npx skills add https://github.com/Acosmi/CrabClaw --skill agent-fleet

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

三省六部军团管理技能提供对由工厂创建的子智能体实例的运行时生命周期的监控、预算治理、熔断保护以及卡死检测,帮助团队在多智能体协同环境中保持可控与安全。

Core Features & Use Cases

  • 实例生命周期监控与状态查询,支持 running、completed、failed、stalled、terminated 等状态转移的可观测性。
  • 实施三层预算体系与熔断保护,确保资源按策略分配并在高风险场景自动降级或阻断。
  • 基于心跳/TTL 的卡死检测,以及自愈与人工干预的升级链路,确保系统稳定性与可持续运行。
  • 提供 RPC 风格的操作与仪表板指标查询,方便运营与审计。

Quick Start

通过 agent-fleet skill 检索当前军团状态并应用默认治理策略。

Frequently Asked Questions about agent-fleet

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

FAQPage Schema
How do I monitor lifecycle states for multiple concurrent agents?

Multi-agent lifecycle monitoring tracks state transitions like running, completed, failed, stalled, and terminated. It exposes RPC-like operations for status queries, metrics, and audit trails, ensuring fleet observability and operational safety.

What is the best way to prevent runaway costs in a multi-agent fleet?

To prevent runaway costs in a multi-agent fleet, apply a three-layer budget system. This governance framework enforces resource allocation policies and automatically triggers circuit-breaker logic to halt operations when high-risk thresholds are breached.

How does heartbeat-based stall detection work for sub-agents?

Heartbeat-based stall detection uses TTL signals to identify stalled sub-agents. When a heartbeat fails, the system triggers an escalation pathway for automated self-healing or manual intervention, ensuring sustainable multi-agent fleet stability.

Can I apply circuit-breaker logic to protect concurrent agent workflows?

Yes, you can apply circuit-breaker logic to protect concurrent agent workflows. This mechanism automatically degrades or blocks operations in high-risk scenarios, preventing cascading failures across the multi-agent fleet while maintaining operational safety.

Does multi-agent fleet management require specific factory-created instances?

Multi-agent fleet management is designed for factory-created sub-agent instances. It provides runtime lifecycle monitoring, budget governance, and fault protection specifically tailored for operational scenarios where multiple sub-agents run concurrently.