Introduction to Skills | Bondar Academy

Lesson Summary

Overview of Skills in AI Agents In this lesson, we explore skills , a form of agent knowledge base that enhances agent capabilities. Skills are recognized as an industry-standard method for extending these capabilities, distinct from rules and cloud.md . What are Skills? Skills are a lightweight , open format designed for extending AI agent capabilities with specialized knowledge and workflows . They differ from other knowledge files by providing not only knowledge but also the procedures to execute tasks. Knowledge vs. Workflows Knowledge: Skills can serve as knowledge bases, but they are invoked dynamically based on the ongoing discussion, unlike cloud.md and rules , which load automatically. Workflows: Skills can encapsulate step-by-step instructions for tasks. For example, a skill can automate the process of creating a branch, running tests, and creating a pull request. Invocation of Skills Skills can be invoked using shortcut commands, allowing for flexibility in their use. In contrast, rules cannot be invoked directly in this manner. Structure and Naming Convention To create a skill, follow these guidelines: Create a folder named after the skill. Inside this folder, create a file named skill.md (with a capital 'S'). Include the relevant workflows or knowledge within this file. In summary, skills are essential for automating workflows in AI agents, providing a flexible and structured approach to enhancing agent functionality. Further details on rules and skills will be covered in subsequent lessons.