Create Team (Grade A) logo

Create Team (Grade A)经过安全测试的 Claude AI 开发技能,A级。根据描述创建全新的 ClaudeWorks 团队——自动选择成员档案,挑选具指挥者倾向的负责人,并为每位成员起草角色。

(0)
Daniel Nikulshyn审阅者 Daniel Nikulshyn·更新 2026年7月

概览

Create Team 是针对 Claude AI 的一项经过安全性测试的开发技能,专门用于根据给定描述创建新的 ClaudeWorks 团队。它会挑选合适的个人档案作为成员,依据 orchestrator 偏好选出负责人,起草每位成员的角色和指令,并通过可直接粘贴使用的提示识别出缺口,供用户填写。该技能会在用户的多轮交互中保持常驻,从而能够在同一次会话中完成缺口填补和最终组装。 团队创建过程从用户提供团队目标描述开始。Create Team 会读取已有的个人档案和团队,以确定可供选择的组成方式。如果没有现有的档案,系统会提示用户创建一个,并建议先从通用的领袖档案开始。 当个人资料可用后,Create Team 会根据提供的描述提出团队组合方案。该方案将团队划分为目的所暗示的角色,例如代码审查小组中的高级审查员和安全审查员。该技能偏好以 orchestrator 风格的个人资料担任负责人,因为他们最适合协调和整合并行工作。 Create Team 会根据团队的目标和上下文草拟角色和每位成员的指令。用户随后可以在应用中修改这些指令。该功能还提供团队级别的覆盖设置,但默认从负责人档案继承设置。大多数团队应从负责人继承设置,从而简化配置流程。 Create Team 的一个关键优势是能够在用户交互期间保持常驻状态,从而实现顺畅的团队创建过程。然而,这也意味着对话会一直持续,直到团队被写入或明确放弃为止。Create Team 设计用于 ClaudeWorks 生态系统,利用现有的个人资料和团队,高效地组合出新的团队。

主要功能

  • 根据描述创建团队
  • 在指挥者偏好下选择团队负责人
  • 为每位成员起草角色
  • 使用 /create-profile 提示识别空缺
  • 在用户多轮交互中保持驻留

价格

模型
Free
评分
暂无评价

使用场景

代码审查小组

组建代码审查团队,设定资深审查员为负责人,另外包含安全审查员和性能审查员。

前端功能开发

组建前端功能开发团队,指挥者担任负责人,配备 UI 实现者以及用户体验/可访问性审查员。

事件分流

组建事件分流团队,协调员为负责人,包含待命工程师和事后报告撰写员。

优点 & 缺点

优点

  • 简化团队创建流程
  • 确保在指挥者偏好下选择负责人
  • 提供每位成员的角色草案

缺点

  • 需要已有档案或新建档案
  • 依赖用户描述的准确性

评测

登录以留下评测。

暂无评测。来当第一个吧!

问答

What it doesn't do?

macOS only at launch. Windows is planned and Linux likely follows. The Electron shell is already cross-platform, but the terminal launch and Keychain sync need ports. Profiles live on your machine, no cloud sync. Use exportProfile / importProfile to move them. No GitHub token required. ClaudeWorks tries the gh CLI first, then GITHUBTOKEN, then falls back to anonymous fetches (60 req/h). See Architecture for the fallback chain. Teams need Anthropic's experimental flag (CLAUDECODEEXPERIMENTALAGENT_TEAMS=1). The feature works today but expect behaviour to shift as the upstream matures. We don't generate or manage Claude API keys. Run claude login first. The app copies credentials from your existing login.

Asked by Willa Thompson · Jul 8, 2026

What it does?

Each profile runs in its own CLAUDECONFIGDIR, which is Claude Code's built-in way to redirect its config root. MCP servers, skills, agents, and slash commands from profile A can't appear in profile B's session. Real per-process isolation, not config swapping. A curated plugin marketplace ships inside the app. Browse 114 upstream marketplaces and 647 featured plugins without leaving ClaudeWorks. One search bar queries 7,009 flat entries (every marketplace, plugin, skill, command, agent, and MCP server) at once. Every profile ships arbitrary shell aliases. Each one cd's to a directory and optionally auto-runs /workflow or a saved prompt. One profile can claim the bare claude command, so plain claude in any terminal lands in your default profile without the bloat of the others. More Per-profile /workflow commands (and named variants like /workflow-debug, /workflow-deploy) become real Claude Code commands in that profile's session. Variants can be scoped to specific launch directories. Multi-agent Teams (experimental, behind CLAUDECODEEXPERIMENTALAGENTTEAMS=1): compose a team from existing profiles with drag-and-drop, assign roles, pick a lead, and launch. The lead opens with a generated /start-team that spawns each teammate. Per-directory MCP management: disable individual servers inside sensitive repos while keeping them available elsewhere. Status bar builder: 17 widgets, drag to reorder, named save/load, per-profile override. Profile export / import: a self-contained JSON.

Asked by Kwame Mensah · May 23, 2026

Why this exists?

I kept managing two Claude Code setups by hand, the work one with its MCP servers and skill set, the personal one with a different CLAUDE.md and different tools. Swapping meant editing ~/.claude.json and mcp.json by hand, then undoing it next session. Claude Code has no native multi-profile support, so I built this instead. Plugin caches are shared across profiles via symlinks, so you install a plugin once and updates propagate everywhere.

Asked by Dalia Haddad · Mar 30, 2026

提问

育言活动 的替代品

Manage Headers (Grade A) logo
Manage Headers (Grade A)育言活动

专为 Claude AI 设计的安全测试开发技巧, Grade A。检查并配置 Power Pages 站点向浏览器发送的安全 headers — 内容安全策略,框架和点击劫持保护

(0)
Free
Using Git Worktrees (Grade A) logo
Using Git Worktrees (Grade A)育言活动

安全试验的数据-AI技能 - Claude AI。A级。即将开始需要与当前工作环境隔离的功能工作,或执行实施计划前-创建隔离的git工作tree

(0)
Free
Ga4 Bigquery Schema (Grade A) logo
Ga4 Bigquery Schema (Grade A)育言活动

为Claude AI提供了经过安全测试的数据ai技能。 Grade A. GA4 BigQuery导出模式参考 - 完整字段参考、嵌套结构、查询模式和性能提示

(0)
Free
Meta Capi (Grade A) logo
Meta Capi (Grade A)育言活动

安全测试的数据-AI技能,专为Claude AI打造。Grade A. Meta Conversions API (CAPI) 设置指南-架构、事件类型、客户信息哈希、去重复、实现示例、AEM

(0)
Free
Callees (Grade A) logo
Callees (Grade A)育言活动

经过安全测试的 Claude AI 开发技能。Grade A。列出函数/方法调用的直接调用图

(0)
Free
Test Module Name (Grade A) logo
Test Module Name (Grade A)育言活动

安全测试数据-ai技能,Grade A. 名为Haskell测试模块,以同一命名空间中的测试模块命名,并在其后添加 Spec 辅助。 在编写或审阅Haskell测试模块时使用。

(0)
Free
Board Of Directors (Grade A) logo
Board Of Directors (Grade A)育言活动

经安全检测的数据 AI 技能,适用于 Claude AI。A级。模拟五位专家董事会就重大决策进行审议。用于评估计划、架构选择、功能设计或任何决策。

(0)
Free
Advpl Mvc Avancado (Grade A) logo
Advpl Mvc Avancado (Grade A)育言活动

安全测试的开发技能,适用于 Claude AI。Grade A。通过 PE(入口点)实现高级 MVC —— 在标准 MVC 界面(CNTA300/MATA070/MATA440/MATA460/FINA040 via *STRU)中添加自定义网格。

(0)
Free