Unverified50% confidenceFactExact time
Spark 采用主从（Master-Worker）架构，Master 负责协调与资源分配，Worker 负责实际计算
1
Sources
50%
Confidence
Long-term
Relevance
8/15/2026
First Seen
Sources
Related Claims
UnverifiedWorkBuddy通过团长中枢角色解决多Agent任务编排问题，团长负责理解用户意图、拆解子任务、分配给对应专家Agent并汇总结果64% similarUnverifiedMuse Spark 1.1的核心价值主张包括处理大型智能体工作负载、自动修复Bug和辅助大规模代码迁移63% similarUnverifiedManifest Studio的Skill机制承担组织目录结构和调用发布接口两项职责，使模型只需识别用户发布意图即可触发60% similarUnverifiedWorkBuddy开发专属Skill的三步法为：安装Skill Creator、用大白话描述需求、确认并完成创建60% similarUnverifiedLight-Skills的Orchestrator组件采用Orchestrator-Worker架构，由编排器负责调度、多个Worker负责执行60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/753286API
curl https://kongchang.com/api/v1/knowledge/claims/753286MCP
get_claim(id=753286)