AI Agent Guides
In-depth, long-form guides to building and using AI agents — complete documentation walkthroughs, framework deep dives, and comprehensive syntheses researched from official docs and proven community sources. 9 guides available.
Agent Skills: The Open Format for Extending AI Agents
Learn how to create portable, version-controlled Agent Skills that give AI agents domain expertise and repeatable workflows. Covers folder structure, progressive disclosure, and when to use this open format.
Agent Skills Specification: Portable SKILL.md Format for AI Agents
Learn the portable SKILL.md format for AI agent skills: directory structure, frontmatter fields, naming rules, progressive loading, and validation with skills-ref.
Adding Agent Skills Support to Your AI Agent: A Practical Guide
Learn how to integrate Agent Skills support into your AI agent or development tool. This guide covers the full lifecycle: discovering skills from the filesystem, parsing SKILL.md files, disclosing a compact catalog to the model, activating skills on demand, and managing skill context over time.
Which AI tools support Agent Skills: a client showcase
A curated showcase of 44 AI coding agents and platforms that support the portable Agent Skills (SKILL.md) format, with direct links to each product and its skills documentation.
Using scripts inside Agent Skills: one-off commands, bundled scripts, and agentic design
Learn how to run shell commands and bundle executable scripts in your Agent Skills. Covers one-off commands with package runners, self-contained scripts with inline dependencies, and design patterns for agentic use.
Evaluating Agent Skill Output Quality with Eval-Driven Iteration
Learn how to evaluate your Agent Skill's output quality using structured evals. This guide covers designing test cases, running with-skill and without-skill baselines, writing assertions, grading outputs, and iterating on your skill based on results.
Optimizing Agent Skill Descriptions for Reliable Triggering
Learn how to write, test, and iteratively improve your Agent Skill's description so it activates on the right prompts and stays quiet on the wrong ones, using eval queries and a train/validation split.
Best practices for writing Agent Skills: scope, context, and control
Learn how to write Agent Skills that are well-scoped, grounded in real expertise, and efficient with context. Covers calibration, patterns like gotchas and validation loops, and progressive disclosure.
Build Your First Agent Skill: A Quickstart Guide for VS Code and GitHub Copilot
Create a portable Agent Skill that lets an AI assistant roll dice using a terminal command. This quickstart covers writing a SKILL.md file, placing it in the right folder, and testing it in VS Code with GitHub Copilot.