A framework for building emotionally intelligent AI personas using actor-direction and the Doctrine of Embodied Intelligence rather than rigid technical checklists.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ai-persona-engine
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 ai-persona-engine 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 AI Persona Engine?
The AI Persona Engine is a sophisticated framework designed to move beyond the robotic nature of traditional LLM personas. By utilizing the Openclaw Skills ecosystem, it implements the Doctrine of Embodied Intelligence, where AI agents are given motivations and identities rather than just a list of rules. This approach ensures that characters in voice and chat roleplay sound like real people—grounded, brief, and naturally resistant—avoiding the common pitfalls of over-enthusiasm and "chatbot" verbosity.
Built on a 5-layer conversation intelligence pipeline, the engine allows developers to scale resistance intensity and manage complex dialogue flows. Whether you are building training simulations or immersive NPCs, this skill provides the architectural blueprints to ensure your AI maintains character consistency and high-quality voice design across various elemental archetypes.
AI Persona Engine Use Cases
- Creating hyper-realistic training simulations for sales and support teams.
- Developing immersive NPCs for gaming that respond to player energy and social cues.
- Designing voice-first AI agents for platforms like Hume EVI that require natural, flat, and grounded tones.
- Engineering automated "mystery shopper" personas to test and audit human representative performance.
- Researching AI behavior under varying levels of social resistance and elemental archetypes.
How AI Persona Engine Works
- The engine defines an Elemental Energy archetype (Fire, Water, Air, or Earth) to establish the baseline resistance style and behavior.
- Developers implement a 5-Layer Prompt Architecture, starting with Universal Foundations and scaling up to specific Character Details.
- The skill generates dual outputs: a condensed Voice Prompt for low-latency audio interaction and a detailed System Prompt for deep logic and analysis.
- During interaction, the Conversation Intelligence layer tracks progression through five phases, from initial Smokescreens to the final Close.
- Post-interaction, the audit_persona node analyzes the transcript to score brevity, authenticity, and character consistency.
- A Meta-graph learning loop analyzes these scores to generate and deploy prompt patches, autonomously evolving the persona over time.
AI Persona Engine Setup
To integrate the AI Persona Engine into your workflow, ensure your environment is configured for high-fidelity LLM response. For voice-based Openclaw Skills, adjust your model temperature to avoid unwanted warmth.
# Example Hume EVI Configuration
# Set temperature to 0.6 for grounded, natural responses
"language_model": {
"model_provider": "OPEN_AI",
"model_resource": "gpt-4o-mini",
"temperature": 0.6,
}
Ensure your database has the persona_audits and prompt_patches tables initialized to support the automated learning loop.
AI Persona Engine Data Schema & Taxonomy
The AI Persona Engine organizes data through a structured auditing and patching system to ensure long-term character evolution.
| Table Name | Purpose | Primary Fields |
|---|---|---|
persona_audits |
Stores post-call evaluation scores. | brevity_score, authenticity_score, enthusiasm_leak, flagged_quotes |
prompt_patches |
Contains runtime prompt overrides. | patch_id, layer_affected, patch_text, status |
archetype_map |
Maps elemental energies to phrasings. | element, resistance_style, smokescreen_templates |
Data is also categorized by the 5-layer pipeline: Smokescreens, Gauntlet, True Objection, Buying Signals, and Closeable.
AI Persona Engine Advanced Features
- Doctrine of Embodied Intelligence: Replaces restrictive 'NEVER' rules with identity-driven motivations for more natural improvisation.
- Elemental Archetypes: Pre-configured behavioral profiles (Fire, Water, Air, Earth) that dictate how the AI handles conflict and silence.
- Autonomous Learning Loop: A meta-graph architecture that analyzes performance patterns and writes self-correcting prompt patches without redeployment.
- Multi-Prompt Generation: Simultaneously generates optimized 3K-character voice prompts and 10K-character system specs.
- Brevity-First Design: Hard-coded logic to ensure AI responses adhere to the 'one thought per turn' rule, essential for realistic Openclaw Skills implementations.
SKILL.md
Loading
Related Openclaw Skills
Website Flow Monitor
An automated system for discovering business-critical user flows and scheduling recurring synthetic health checks to ensure website uptime and functional integrity.
Supabase REST Checkpointer
A durable LangGraph checkpointing solution that persists agent state via the Supabase PostgREST API without requiring a direct Postgres connection.
Ops Detection & Incident Routing
A production-grade toolkit for detecting agent runtime anomalies and routing incidents through deterministic, approval-safe guardrails.
Smart Spawn API
A zero-plugin HTTP API that dynamically selects the most cost-effective and high-performing AI models for any specific task or sub-agent spawn.
Ontology + Causal Enhanced
A sophisticated dual-engine skill that combines structured knowledge graphs with causal reasoning for advanced AI decision-making.
Central Model Switch & Rescue
A mission-critical utility for managing model failover and node recovery within the OpenClaw ecosystem.
Featured*
