What files does this skill manage?

It manages the core OpenClaw workspace files: SOUL.md (personality), IDENTITY.md (branding), AGENTS.md (instructions), MEMORY.md (knowledge), and GOALS.md (objectives).

Can I use this on an existing OpenClaw setup?

Yes, the skill can scan existing workspaces to identify inconsistencies, redundant instructions, or areas where the persona can be tightened for better performance.

Why is token budget mentioned?

Since files like SOUL.md are loaded into every prompt, keeping them concise is critical to saving tokens and maintaining model focus; the skill helps optimize this balance.

How does the interview process work?

The skill conducts a phased interview asking targeted questions about your agent's role, values, and style before generating the corresponding markdown files.

OpenClaw Agent Architect

Name: OpenClaw Agent Architect
Author: TheophilusChinomona

byTheophilusChinomona

0•

Productivity & Workflow

Architects and optimizes OpenClaw AI agent personas and operational frameworks through guided workspace management.

The OpenClaw Agent Architect skill is a specialized tool for Claude Code designed to build, refine, and audit the foundational markdown files that define an OpenClaw agent's personality and behavior. By leveraging a structured interview process, it generates production-ready SOUL, IDENTITY, AGENTS, MEMORY, and GOALS files, ensuring a cohesive and efficient agent workspace. Whether you are bootstrapping a new agent from scratch or auditing an existing configuration for token efficiency and behavioral consistency, this skill provides the domain-specific logic needed to create sophisticated, self-hosted AI assistants.

Key Features

01Token-budget awareness to ensure agent instructions remain lean and performant

02Workspace auditing to detect bloat, vagueness, or inconsistencies in existing configurations

03Integrated management of memory structures and goal-tracking frameworks

040 GitHub stars

05Automatic generation of optimized SOUL.md, IDENTITY.md, and AGENTS.md files

06Guided phase-based interviews for agent personality and identity creation

Use Cases

01Building a new specialized AI agent with a unique persona and operational rules

02Auditing an existing OpenClaw workspace to improve response quality and reduce token costs

03Standardizing agent behaviors and safety guardrails across a self-hosted AI team

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add theophiluschinomona/claude-openclaw-plugin openclaw-agent-architect

For use in Claude.ai and ChatGPT

Key Features

01Token-budget awareness to ensure agent instructions remain lean and performant

02Workspace auditing to detect bloat, vagueness, or inconsistencies in existing configurations

03Integrated management of memory structures and goal-tracking frameworks

040 GitHub stars

05Automatic generation of optimized SOUL.md, IDENTITY.md, and AGENTS.md files

06Guided phase-based interviews for agent personality and identity creation

Use Cases

01Building a new specialized AI agent with a unique persona and operational rules

02Auditing an existing OpenClaw workspace to improve response quality and reduce token costs

03Standardizing agent behaviors and safety guardrails across a self-hosted AI team

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add theophiluschinomona/claude-openclaw-plugin openclaw-agent-architect

For use in Claude.ai and ChatGPT