Roundtable Pai for Openclaw

A multi-turn dialogue skill that orchestrates simulated roundtable discussions between three expert personas to provide deep, multi-perspective insights.

zhouxiang1
v1.6.5
Apr 17, 2026
0
990
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install roundtable-pai

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install roundtable-pai 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 Roundtable Pai?

Roundtable Pai is a sophisticated multi-turn conversation framework designed for Openclaw Skills. Unlike standard one-shot generators, this skill facilitates a structured debate between three distinct simulated personas chosen from an expert registry. It uses a dedicated state-management controller to ensure continuity and logical progression across multiple interaction rounds, allowing users to explore complex topics through a simulated panel of masters.

The skill is built on a foundation of safety and state persistence, relying on a local JSON runtime to track the conversation rather than just model memory. This ensures that every character stays in persona and every discussion follows a rigorous, high-quality analytical framework.

Roundtable Pai Use Cases

  • Simulating expert panels to analyze philosophical, technical, or strategic questions.
  • Exploring diverse viewpoints on a single topic through interactive multi-turn dialogue.
  • Generating high-quality character-driven content based on historical or professional personas.
  • Structured brainstorming where the user can intervene, agree, or steer the direction of a multi-agent debate.

How Roundtable Pai Works

  1. The user initiates the skill by posing a question or using a trigger phrase like start roundtable.
  2. The system executes a controller script to pull a pool of experts and asks the user to select three participants.
  3. Once personas are selected, the skill initializes the first round of discussion, referencing persona-specific data and discussion frames.
  4. After each round, the user is presented with options (A/B/C/D/E) to agree with a specific expert, stay silent, or contribute their own input.
  5. The controller script processes each turn, updating the state in a local JSON file to maintain consistency.
  6. Upon reaching the round limit or a natural conclusion, the skill generates a final summary across six structured fields.

Roundtable Pai Setup

To deploy Roundtable Pai within your Openclaw Skills environment, ensure the controller script and data directories are correctly mapped.

# Example of manually triggering the controller for state initialization
python3 ${CLAUDE_SKILL_DIR}/scripts/roundtable_controller.py --stdin <<'EOF'
Should silicon-based intelligence eventually replace biological life?
EOF

Ensure that the runtime/ directory is writable for the skill to maintain the roundtable_state.json file, and that the references/ folder contains the necessary persona and frame markdown files.

Roundtable Pai Data Schema & Taxonomy

The skill organizes its operations through a strict directory and file structure to ensure isolation and data integrity:

File/Path Purpose
data/character_registry.json The master list of available expert personas and their metadata.
runtime/roundtable_state.json Tracks the current round, selected characters, and the history of the debate.
references/personas/*.md Detailed behavioral profiles and linguistic styles for each expert.
references/discussion-frames/*.md Templates that define the logical flow and quality standards of the debate.

Roundtable Pai Advanced Features

  • State-persistent multi-turn logic that prevents context drift during long discussions.
  • Character-driven simulation using specialized linguistic modeling for distinct expert voices.
  • Structured user interaction hooks (A/B/C/D/E) for precise control over the simulation flow.
  • Strict security auditing that utilizes heredoc stdin to prevent shell injection and path traversal attacks.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*