When you've got a complex task that can be split into parallel workstreams, this spins up a team of Claude instances with different roles and models. It walks you through designing the team structure (analyst on Opus, implementer on Sonnet, etc.), gets your approval, then uses TeamCreate to run specialists in parallel or sequence based on dependencies. The workflow is solid: analyze, propose roles, execute with proper blocking, aggregate results, clean up. Most useful for things like research plus implementation plus validation, or any job where you'd naturally want multiple experts working simultaneously. The built-in role to model mapping prevents you from burning Opus tokens on grunt work.
npx -y skills add ai-native-camp/camp-2 --skill team-assemble --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sickn33/antigravity-awesome-skills
moizibnyousaf/ai-agent-skills
github/awesome-copilot