AI agents are only as good as their tools and instructions. This guide helps you design clear action spaces, tool definitions, and observation formats. The goal is to get more tasks completed successfully.
You will learn how to name tools, structure their inputs, and write helpful output messages. The guide also covers error recovery with simple steps and hints. A steady system prompt and compact observations keep the agent focused.
Different tasks need different approaches. Exploratory jobs benefit from ReAct planning. Structured flows work best with function calling. A hybrid pattern combines both for reliable results.
Global
mkdir -p ~/.claude/skills/agent-harness-constructionProject
mkdir -p .claude/skills/agent-harness-constructionSource Repository
Find Skillsvercel-labs/skills
Find and install the perfect skill to extend your AI agent
Microsoft Foundrymicrosoft/azure-skills
Build, deploy, and improve AI agents on Microsoft Foundry from start to finish
Azure Aimicrosoft/azure-skills
Search, transcribe, and analyze with Azure AI tools for smarter apps
Azure Hosted Copilot Sdkmicrosoft/azure-skills
Build, deploy, and manage your Copilot SDK apps on Azure with ease
Skill Creatoranthropics/skills
Create, test, and improve AI agent skills with easy step-by-step guidance
Image Editagentspace-so/runcomfy-agent-skills
Smart router picks the best AI model for your image editing needs
Agentspaceagentspace-so/skills
See your AI agent's live folder from any browser instantly
Openclaw Secure Linux Cloudxixu-me/skills
Secure self-hosting of OpenClaw with private control and SSH tunneling