council

What This Skill Does

The Council of the Wise is a high-level orchestration skill for OpenClaw that leverages multi-agent collaboration to stress-test your ideas. By scanning your local agents directory, the skill dynamically builds a panel of experts to review your proposals, business plans, or technical designs. It combines distinct domain-specific personas—ranging from technical architects to creative artists—to provide a balanced, critical, and nuanced analysis of any topic you submit.

Installation

To integrate this skill into your environment, run the following command in your terminal: clawhub install openclaw/skills/skills/jeffaf/council-of-the-wise Ensure you have a populated agents/ folder to take full advantage of the auto-discovery features.

Use Cases

• Strategic Planning: Pitch a new business model and have the Analyst and Architect identify hidden costs and structural flaws.
• Technical Design: Present a software architecture to the Engineer and Pentester to catch security vulnerabilities and implementation bottlenecks.
• Content & Creative: Refine marketing copy or creative stories by gathering feedback on voice, tone, and user experience.
• Complex Decision Making: When you are at a crossroads, use the Council to force a confrontation between different logical frameworks.

Example Prompts

1. "Send this to the council: I am planning to pivot my startup to a subscription-only model for our core API, what are the risks?"
2. "Council of the wise: Evaluate this project proposal for an automated document filing system."
3. "Get the council's feedback on my current plan to rewrite the backend in Rust."

Tips & Limitations

• Persona Customization: You can easily influence the Council by adding or removing .md files in the agents/ folder. The skill is as strong as the personas you define.
• Synthesis Focus: Always pay attention to the Synthesis section at the top of the output; it is designed to resolve conflicts between disagreeing members.
• Patience: Because this skill spawns multiple sub-agents to provide deep analysis, expect a processing time of 2-5 minutes. It is a tool for deep work, not instant messaging.
• Confidence Signals: Look for the confidence score at the bottom of the output. Low confidence indicates that your idea needs further research or has significant unresolved contradictions.