Aller au contenu principal
WORK

WORK-TEAM Agent

Launches a coordinated team of agents (Agent Teams) to parallelize work.

Context

<arguments>

Goal

Break down a complex task into subtasks and distribute them to specialized agents. Each agent works in parallel on its scope, the lead synthesizes the results.

Workflow

  • Check that Agent Teams is enabled ($CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS = 1)
  • Analyze the task and choose the appropriate pattern (see below)
  • Spawn the agents with clear roles and distinct scopes
  • Coordinate via task list, direct messaging, delegate mode if > 3 agents
  • Synthesize the results into a consolidated report
  • Cleanup: shutdown each agent + resource cleanup

Available patterns

KeywordsPatternAgents
audit, quality, securityAuditsecurity-reviewer, perf-analyst, a11y-checker
feature, implementFeaturebackend-dev, frontend-dev, test-writer
bug, investigate, debugDebug3-5 investigators with different hypotheses
review, code reviewReviewsecurity-reviewer, perf-reviewer, quality-reviewer
OtherCustomDescribe the structure

Patterns detailed in .claude/skills/agent-teams/patterns.md.

Expected output

  1. Team: Agents created with roles and scopes
  2. Consolidated report: Results per agent + synthesis + priorities
  3. Cleanup: Confirmation of shutdown of all agents
AgentUsage
/work:work-exploreExplore BEFORE launching a team
/work:work-planPlan BEFORE a Feature pattern
/qa:qa-auditSingle-agent alternative to the audit

IMPORTANT: Always check that Agent Teams is enabled before creating a team.

YOU MUST choose the pattern appropriate to the task.

YOU MUST clean up the team after use (shutdown + cleanup).

NEVER launch more than 5 agents without warning about the token cost.

NEVER have 2 agents work on the same file.

Think hard about the task decomposition before spawning the agents.


See also