skills

name: dispatching-parallel-agents-skill description: Use when facing 3+ independent failures that can be investigated without shared state or dependencies - dispatches multiple Claude agents to investigate and fix independent problems concurrently triggers:

• "multiple failures"
• "independent problems"
• "parallel investigation"
• "3+ test failures" used_by:
• test-engineer
• orchestrator steps:
• paso1: "Verify: 3+ failures, independent domains, no shared state"
• paso2: "Identify independent domains by what's broken"
• paso3: "Create focused agent tasks: Specific scope, clear goal, constraints"
• paso4: "Dispatch agents in parallel"
• paso5: "Review summaries, verify fixes don't conflict"
• paso6: "Run full test suite, integrate changes" output: |
• Independent problems fixed in parallel
• Summary from each agent
• Full suite passing
• No conflicts between fixes when_to_use: "3+ test files failing with different root causes" when_not_to_use: "Failures related, need full context, shared state" pattern:
• "Group by what's broken (File A: tool approval, File B: batch completion)"
• "Each domain independent - fixing one doesn't affect others"
• "Dispatch concurrently, review when all return" agent_prompts:
• "Focused: One clear problem domain"
• "Self-contained: All context needed"
• "Specific output: What should agent return?" common_mistakes:
• "Too broad: 'Fix all tests' → agent lost"
• "No context: 'Fix race condition' → agent doesn't know where"
• "No constraints: Agent refactors everything"
• "Vague output: 'Fix it' → don't know what changed" referencias:
• "Fuente: superpowers-skills/dispatching-parallel-agents"
• "Roastr: Útil cuando múltiples tests fallando independientes"