You don't always need the full pipeline. If you already know the options, skip straight to /converge for a structured debate.
Your team is split on state management for a new React app: Redux Toolkit vs Zustand vs React Query + Context. Everyone has opinions. You want a structured way to evaluate the trade-offs.
/converge 3 "We need to pick a state management approach for our new React dashboard. Options: (1) Redux Toolkit — the team knows it, but it's verbose. (2) Zustand — minimal boilerplate, but less ecosystem. (3) React Query + Context — server state in RQ, minimal client state in Context."
What happens:
- Claude frames the debate with evaluation criteria (learning curve, boilerplate, performance, ecosystem, testing)
- Three teammates are assigned:
- Redux advocate: argues for familiarity, middleware ecosystem, devtools
- Zustand advocate: argues for simplicity, bundle size, less boilerplate
- RQ + Context advocate: argues for separation of concerns, cache management
- Two rounds of debate where each advocate steel-mans their position AND concedes genuine weaknesses
- Synthesis with resolved points, remaining tensions, and a recommendation
- All three approaches handle the core requirements. This is a trade-off decision, not a capability decision.
- Server state should be managed by React Query regardless of choice (Redux advocate conceded this in Round 2).
- Familiarity vs learning curve: Redux has team knowledge but Zustand is learnable in a day. Trade-off is real but small.
- Ecosystem vs simplicity: Redux middleware ecosystem matters IF you need sagas/thunks for complex flows. If most state is server-derived, it's unnecessary complexity.
Zustand for client state + React Query for server state. The team's Redux knowledge transfers easily, the dashboard is primarily server-data-driven (favoring RQ), and Zustand handles the small amount of client state without boilerplate overhead.
- Architecture decisions with 2-4 clear options
- Technology choices where the team has differing preferences
- Design trade-offs where both sides have merit
- Post-mortem analysis where you want to evaluate what went wrong from multiple angles
The key insight: /converge forces steel-manning. Each advocate must acknowledge what the other positions get right, which produces better analysis than a simple pros/cons list.