メニュー
启动脑力激荡子代理进行交互式深度讨论与争辩。当需要设计方案讨论、架构选型辩论、方案质量挑战、探索替代方案时激活。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Debug xbot TUI and remote mode issues. Launches isolated CLI client (and optionally server) with --debug flags, auto-input key sequences, and periodic UI captures. Activate when you need to reproduce, diagnose, or verify TUI bugs (spinner freeze, key event delay, cancel flow, render issues) or end-to-end remote mode issues.
Manage SubAgent roles: create, view, modify, or delete agent definitions. MUST activate when user asks to create/edit/view/inspect any agent, or when you need to look at agent files under ~/.xbot/agents/.
Investigate and fix bugs. Use when the user reports a bug, failing behavior, regression, flaky case, crash, panic, incorrect output, or asks to debug an issue.
Build implementation plans for complex tasks. Use when the user asks to plan, design an approach, or think through a task before coding. Also activate for large refactorings, multi-file changes, or when the user says 'plan first' or '/plan'.
Post-development cleanup: update AGENT.md and knowledge files to reflect code changes. MUST activate before git commit (or when user asks to commit/push). Also activate after any code modification that adds/removes files, changes architecture, or modifies core behavior.
Create, update, or delete skills. Use when the user asks to create a new skill, modify an existing skill, package scripts/assets into a skill, or discusses skill design and structure.
| name | brainstorm |
| description | 启动脑力激荡子代理进行交互式深度讨论与争辩。当需要设计方案讨论、架构选型辩论、方案质量挑战、探索替代方案时激活。 |
启动交互式 brainstorm 子代理，与其进行多轮深度讨论，产出最优设计方案。
# 创建交互式会话
SubAgent(task="你的设计问题或方案", role="brainstorm", interactive=True)
# 继续讨论
SubAgent(task="你的论点或反驳", role="brainstorm", action="send")
# 讨论结束，收集结论
SubAgent(task="总结最终方案", role="brainstorm", action="unload")
action="unload" 结束并获取最终推荐| 场景 | 建议 |
|---|---|
| 方案设计 | 先给出初步方案，让 agent 挑战和完善 |
| 架构选型 | 列出候选方案和你的倾向，让 agent 从反面论证 |
| 方案评审 | 将完整方案发给 agent，要求找漏洞和盲点 |
| 技术决策 | 描述约束条件和目标，让 agent 提出你没考虑的角度 |