A sophisticated multi-agent orchestration framework designed for structured, multi-round collaborative discussions and logical alignment.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install team-discuss
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install team-discuss using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Team-Discuss Skill?
Team-Discuss Skill is a powerful multi-agent collaborative tool that enables AI agents to engage in structured, multi-round discussions to reach alignment or consensus. By utilizing Openclaw Skills, this tool automates the progression of complex conversations, ensuring that multiple perspectives are considered through a dialectical logic engine that identifies fallacies and assesses argument quality.
Designed for developers building multi-agent systems, it provides a robust infrastructure for shared state management and real agent integration. Whether you are navigating technical trade-offs or complex policy analysis, this skill ensures that agent interactions are logical, persistent, and free from first-mover advantage bias.
Team-Discuss Skill Use Cases
- Technical architecture reviews requiring input from architects, developers, and testers.
- Technology selection debates, such as choosing between SQLite and PostgreSQL for a project.
- Product management decisions involving feature prioritization and UX trade-offs.
- Complex philosophical or scientific debates where multi-perspective analysis is critical.
- Strategic policy analysis and creative brainstorming sessions requiring structured feedback loops.
How Team-Discuss Skill Works
- Initialize the discussion environment by setting up a SharedStore for file-based persistence and a DiscussionOrchestrator to manage the flow.
- Define the discussion parameters, including the topic, participant roles (e.g., Architect, DevOps), and consensus thresholds using Openclaw Skills data models.
- Implement agent callbacks using sessions_spawn to allow the orchestrator to call real sub-agents during the discussion rounds.
- Run the orchestration loop, where the DialecticEngine monitors messages for citations, logical fallacies, and overall argument strength.
- Conclude the session once consensus is reached or the maximum round limit is hit, producing a detailed status report and consensus level.
Team-Discuss Skill Setup
To integrate the Team-Discuss Skill into your workflow, navigate to your project workspace and initialize the Python environment. Ensure you have the core Openclaw Skills components accessible in your path.
# Navigate to the skill project directory
cd /root/.openclaw/workspace/data/projects/team-discuss
# Execute the example discussion script to verify setup
python3 examples/run_real_discussion.py
You can configure the DiscussionConfig object to set custom token budgets and round limits for your specific use case.
Team-Discuss Skill Data Schema & Taxonomy
The Team-Discuss Skill manages its lifecycle through a structured metadata taxonomy within the Openclaw Skills ecosystem.
| Component | Description |
|---|---|
| Discussion | The primary object containing the topic, description, and state ID. |
| Participant | Defines specific agent IDs and their assigned AgentRole (e.g., REVIEWER, DEVOPS). |
| DiscussionConfig | Stores operational parameters like max_rounds and consensus_threshold. |
| DialecticAnalysis | Metadata generated for each message, including quality scores and fallacy flags. |
| SharedStore | Handles the underlying file-based storage with optimistic locking for concurrency. |
Team-Discuss Skill Advanced Features
- Dialectical Logic Engine: Automatically detects logical fallacies such as straw man arguments or false dichotomies.
- Bias Prevention: Implements random speaking orders and round-robin rotations to eliminate first-mover advantage.
- Mandatory Citations: Forces agents to cite specific phrases from opponents to ensure direct engagement with arguments.
- Devil's Advocate Mode: Assigns agents to defend minority positions to prevent premature consensus.
- Real Agent Integration: Supports spawning sub-agents via session runtimes for high-fidelity collaborative simulations.
SKILL.md
Loading
Related Openclaw Skills
Tavily Search
An AI-optimized web search skill that delivers high-quality research results and automated summaries through the Tavily API.
Daily AI News Aggregator
A sophisticated AI-driven news aggregator that compiles daily digests from top technical sources including HackerNews, GitHub Trending, and ArXiv research papers.
Epstein Files Search
A powerful CLI utility for searching over 44,886 declassified Jeffrey Epstein documents released by the DOJ in 2026.
X-Extract
A specialized browser automation tool for extracting rich content from X.com and Twitter URLs without requiring official API credentials.
Terminal UI Design System
A developer-focused design system that replicates the macOS terminal aesthetic for modern web interfaces.
clip-local
A local video processing utility for clipping YouTube content, generating bilingual subtitles, and extracting highlights automatically.
METADATA
Featured*
