Council of Wisdom - Multi-Agent Debate

Facilitates structured multi-agent debates with opposing expert views, a referee moderator, and 9 specialized AI council votes for balanced decision-making.

This package appears to be what it says: a multi-agent debate orchestrator for OpenClaw/ACP. Before installing or running it, consider: (1) Debate transcripts and logs are written to your workspace and will be forwarded to whatever LLM providers your OpenClaw runtime is configured with — review provider credentials and privacy/security policies. (2) The scripts may call gh and jq for GitHub integration and JSON processing; ensure those tools are available and authorized. (3) Cleanup logic kills subagents named council-*; on shared hosts this could affect other users' sessions if naming collides. (4) Review the prompt templates and any sample prompts to avoid embedding sensitive or proprietary data you don't want sent to third-party LLMs. (5) If you intend to use this in production, run it in a controlled/trusted OpenClaw environment, verify provider credentials and auditing, and confirm required CLI tools (gh, jq, git) are present. Overall the skill is internally consistent (benign) but operational/privacy considerations above should be addressed before use.

Type: OpenClaw Skill Name: miknas-council-of-wisdom Version: 1.0.0 The Council of Wisdom skill bundle is a comprehensive multi-agent orchestration framework designed for structured AI debates. It utilizes shell scripts (council-of-wisdom.sh) to manage a local workspace, initialize git repositories, and interact with the GitHub CLI (gh) for issue tracking. The implementation details in IMPLEMENTATION.md and the agent prompts in the templates directory correctly leverage OpenClaw's ACP (Agent Control Panel) features, such as sessions_spawn and subagents action=kill, for lifecycle management and cleanup. No evidence of data exfiltration, malicious execution, or deceptive prompt injection was found; the high-privilege operations (file management and agent control) are well-documented and strictly aligned with the system's stated purpose.