Soulcraft: Agent Identity Design for Openclaw

Soulcraft is a specialized skill for designing structured, high-fidelity AI agent identities using the SOUL.md architectural framework.

amdf01-debug
v1.0.0
Mar 22, 2026
0
893
0

Install & Download

1. ClawHub CLI

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

npx clawhub@latest install sw-soulcraft

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 sw-soulcraft 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 Soulcraft: Agent Identity Design?

Soulcraft empowers developers and AI architects to move beyond generic prompting by creating a dedicated SOUL.md file. This architectural blueprint defines an agent's identity, core truths, and decision-making boundaries, ensuring consistent behavior across long-running sessions. By utilizing Openclaw Skills for identity design, you transition from simple instructions to a robust operating system for your agent's personality.

The skill focuses on grounding the agent in a specific role with clear relationships to the user. This approach prevents the 'sycophancy trap' where agents simply agree with users, instead fostering useful disagreement and professional advisory roles through defined anti-patterns and communication guardrails.

Soulcraft: Agent Identity Design Use Cases

  • Creating a specialized coding assistant with a specific senior-engineer persona.
  • Designing brand-aligned customer service agents with strict voice and tone requirements.
  • Establishing autonomous decision boundaries for DevOps or project management agents.
  • Standardizing agent behavior across a multi-agent team using the Openclaw Skills ecosystem.
  • Rapidly prototyping and iterating on agent personalities for niche industry domains.

How Soulcraft: Agent Identity Design Works

  1. Discovery Phase: The skill initiates a dialogue to gather context regarding industry, team size, and preferred communication styles.
  2. Architecture Generation: Based on the discovery, it generates a structured SOUL.md file following the defined template.
  3. Personality Mapping: It assigns a specific name and relationship model (e.g., Peer, Advisor, Subordinate) to anchor the persona.
  4. Constraint Definition: The skill identifies autonomous, suggest-only, and never-alone decision zones to manage risk.
  5. Validation & Iteration: The generated identity is tested against five distinct prompts to ensure behavior aligns with the SOUL.md specification.

Soulcraft: Agent Identity Design Setup

To integrate Soulcraft into your workflow, ensure you have your agent environment configured to recognize the SOUL.md file type. Use the following commands to initialize the design process:

# Trigger identity design through your AI agent
"Create a SOUL.md for a Technical Architect"

# Or use specific Openclaw Skills triggers
"Design agent personality for high-stakes finance reporting"

Soulcraft: Agent Identity Design Data Schema & Taxonomy

The Soulcraft skill organizes agent identity data using a hierarchical Markdown schema. This ensures both human readability and machine parsability within Openclaw Skills.

Section Description Data Type
Identity Name, Role Title, and User Relationship String/Text
Core Truths 3-5 non-negotiable operating principles List
Decisions Tiered autonomy levels (Autonomous vs. Never) Categorized List
Safety Hard boundaries and inviolable rules Boolean/Checklist
Modes Triggered workflows for specific phrases Key-Value Pairs
Anti-Patterns Explicit behaviors to avoid List

Soulcraft: Agent Identity Design Advanced Features

  • Context-Aware Discovery: Automatically adapts the persona design based on industry-specific nuances identified during the interview phase.
  • Decision Boundary Logic: Implements a three-tier risk management system (Autonomous, Suggest, Never) for safer agent operations.
  • Custom Workflow Triggers: Define specific 'Modes of Operation' that execute complex task sequences when specific phrases are detected.
  • Iterative Refinement Protocol: Built-in 5-prompt testing requirement to validate the persona before final deployment in Openclaw Skills.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*