ollivanders-agent-shop

What This Skill Does

Ollivanders Agent Shop is a specialized architectural skill designed to bridge the gap between abstract character concepts and functional, production-ready AI agents. Unlike standard prompt engineering tools, this skill treats the agent as a "persona-driven organism" that requires a stable cognitive structure, a defined professional identity, and a robust file-system landing. It operates on the core philosophy that an agent's personality should be a functional outcome of its work identity, not merely a decorative veneer. The skill guides you through a rigorous design process—moving from high-level role definition to specific file scaffolding—ensuring that the final agent is consistent, predictable, and perfectly aligned with the tasks it owns.

Installation

To integrate this skill into your environment, use the following OpenClaw command in your terminal: clawhub install openclaw/skills/skills/child2d/ollivanders-agent-shop

Use Cases

• New Agent Creation: Developing a novel agent from a vague character concept into a structured, reliable tool.
• Persona Redesign: Transforming an existing, ineffective agent by re-engineering its core work identity and cognitive framework.
• Professional Alignment: Ensuring that an agent tasked with technical or critical business functions adopts a personality that reinforces its analytical reliability.
• Production Scaffolding: Planning the directory structure, file skeletons, and system-level registration required for an agent to exist as a long-term, maintainable codebase.

Example Prompts

1. "I want to build an agent that acts as a cynical but brilliant senior code reviewer. Can we start designing the persona and workspace?"
2. "My current agent is too generic. Use the Ollivander process to help me redefine its cognitive structure so it stops giving fluff answers and starts focusing on architectural analysis."
3. "I have an idea for a 'Project Archivist' agent that manages my documentation. How should I structure its files to ensure it maintains a consistent tone and keeps my files organized?"

Tips & Limitations

• Design First: The skill is strictly ordered. If you attempt to skip the design phase, the skill will nudge you back to ensure the agent's foundation is sound. Do not rush to generate files until you have confirmed the design.
• Scope: This is not a deployment tool for runtime orchestration. It focuses on the creation, structure, and persona logic of an agent. It does not handle cloud hosting or API keys.
• Cognitive Fidelity: Treat the 'Identity-Binding' sentence as the North Star. If your agent starts hallucinating out of character, revisit this sentence to ensure it provides sufficient guardrails for the agent's work identity.