From helloagents
Orchestrates sub-agents for task execution, parallel development, and work dispatching. Ensures clear context, verification standards, double-stage reviews, and conflict-free collaboration.
npx claudepluginhub hellowind777/helloagents --plugin helloagentsThis skill uses the workspace's default tool permissions.
子代理编排必须遵循以下规范。
Guides creating, configuring, and orchestrating Claude Code subagents and Task tool. Covers prompts, tools, models, file structures, and multi-agent workflows.
Dispatches subagents for multi-task code implementation using parallel execution for independent tasks and two-stage reviews for spec compliance and code quality. Triggers on plan.json tasks or user requests.
Executes multi-task implementation plans in current session via isolated subagents per task: implement, spec review, code quality review. For independent tasks without context switching.
Share bugs, ideas, or general feedback.
子代理编排必须遵循以下规范。
.helloagents/ 在本 skill 中统一按项目级存储路径理解：状态文件只使用 state_path；会话证据使用当前 state_path 所在目录下的 artifacts/*.json；若 project_store_mode=repo-shared，方案包、verify.yaml 与 DESIGN.md 按当前上下文中已注入的项目知识/方案目录解析。
先确定任务是否适合子代理（独立性高、边界清晰、可验证）。
.helloagents/DESIGN.md（按当前项目存储模式解析）或其中相关片段helloagents-turn-state write 代替主代理写完成态或等待态；需要运行时状态信号时只由主代理写入