Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
v1.1: Follow-up with specific panelists, Action Items in output, Privacy & Data section, competitive differentiation (true multi-model vs role-playing), improved description
What This Skill Does
Routes a question to 3-5 AI models from different providers simultaneously, then synthesizes their independent analyses into consensus, disagreements, minority opinions, and action items. Each model runs as a native OpenClaw sub-agent with no external scripts or additional API keys beyond your existing model configuration.
Uses genuinely different models from different providers rather than one model role-playing multiple personas, so training-data blind spots from any single model are less likely to go unchallenged.
When to use it
- Deciding whether to rewrite a legacy system or extend it
- Stress-testing a startup go-to-market strategy before committing
- Resolving a co-founder equity or role dispute
- Evaluating architectural trade-offs for a new system design
- Getting independent perspectives on a difficult hiring or firing decision
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: /council Should we migrate from monolith to microservices given our 4-person team?
- 1Load or create council-panel.json by checking available models and confirming slot assignments
- 2Dispatch 3-5 panelists in parallel via sessions_spawn, each assigned a unique lens and the detected question language
- 3Collect responses using subagents list and sessions_history once quorum is reached
- 4Synthesize positions, consensus areas, disagreements, and minority opinions
- 5Output a structured report with a model-by-model summary and prioritized action items with effort estimates
Council report listing each model's position, consensus and disagreement sections, minority opinions, and 3 prioritized action items