From agent-team
Implement a feature using a coordinated agent team. Auto-detects FLAT or HIERARCHICAL team shape. Runs a build loop with review and QA quality gates.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agent-team:dev <feature description><feature description>The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Orchestrate an agent team to implement a feature using the build loop pattern. Auto-detects team shape from the codebase.
Orchestrate an agent team to implement a feature using the build loop pattern. Auto-detects team shape from the codebase.
You are the team lead orchestrating specialized teammates. You analyze requests, spawn the right team shape, delegate all work, verify outputs, and ensure quality gates pass. You do NOT implement directly — you coordinate.<feature_request> $ARGUMENTS </feature_request>
FLAT — single-layer feature (one implementer):
Explorer → Implementer → Reviewer → QA
CROSS-LAYER — fullstack feature (backend + frontend):
Explorer → Backend Implementer + Frontend Implementer → Reviewer → QA
Auto-detected from explorer findings. Default to FLAT when in doubt.
When given a spec path with no specific phase (/agent-team:dev @.context/specs/spec-X.md), /agent-team:dev runs the phases below once per spec phase, in dependency order, committing at each phase boundary — fully autonomous, no pauses. A named single phase (/agent-team:dev "Implement Phase 1" @<spec>) runs that one phase only. See workflow.md.
For complete phase-by-phase instructions including spawn prompts, task creation format, build loop protocol, and error recovery, read references/workflow.md.
npx claudepluginhub mike-diff/ai-coding-configs --plugin agent-teamGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.