skill markdown must contain YAML front-matter(name and description)

Based on the project documentation, the system identifies each skill by parsing metadata embedded directly within the files or their directory structures. Here is exactly how it works:

1. YAML Front-Matter (The Unique Identifier) Each Markdown skill file (SKILL.md) must contain YAML front-matter at the very top of the document. This front-matter requires a specific format containing two mandatory fields:

  • name: This acts as the unique skill identifier for the agent. Furthermore, the value in the name field must exactly match the name of the directory the skill resides in.
  • description: A human-readable description of what the skill does and when the agent should use it.

Example of a skill file's structure:

---
name: orchardcore.example
description: A description of what this skill does and when to use it.
---

# Skill Title
Skill content goes here...

2. System Parsers The base MCP engine (CrestApps.AgentSkills.Mcp) handles the discovery and extraction of this information using dedicated internal parsers:

  • SkillFrontMatterParser: Specifically built to parse .md files and extract the YAML front-matter metadata.
  • SkillYamlParser: If a skill is defined using a raw YAML file (SKILL.yaml or SKILL.yml) instead of Markdown, this parser extracts the identical name and description identifiers.
  • SkillFileParser: A unified parser that detects the file format and delegates the job to the correct parser above.

Through this combination of standardized front-matter and automated parsers, the MCP engine correctly identifies the name and body of the skill, translating it into an MCP Prompt or Resource for the AI agent to use.