Crate skills_ref 

skills-ref-rs

Rust implementation of the agentskills library for validating, parsing, and managing Agent Skills - structured markdown documents that describe capabilities available to AI assistants.

Core Functionality

1. Validate - Check that a skill directory contains a valid SKILL.md
2. Read Properties - Parse SKILL.md frontmatter and extract metadata as JSON
3. To Prompt - Generate XML blocks for embedding skill information in agent prompts

Example

use skills_ref::{read_properties, validate, to_prompt};
use std::path::Path;

// Validate a skill directory
let errors = validate(Path::new("my-skill"));
if errors.is_empty() {
    println!("Skill is valid!");
}

// Read skill properties
let props = read_properties(Path::new("my-skill")).unwrap();
println!("Skill name: {}", props.name);

// Generate XML prompt
let xml = to_prompt(&[Path::new("my-skill")]).unwrap();
println!("{}", xml);

Re-exports

pub use error::Result;

pub use error::SkillError;

pub use models::SkillProperties;

pub use parser::find_skill_md;

pub use parser::parse_frontmatter;

pub use parser::read_properties;

pub use prompt::to_prompt;

pub use validator::validate;

pub use validator::validate_metadata;

Modules

error

Skill-related error types.

models

Data models for Agent Skills.

parser

YAML frontmatter parsing for SKILL.md files.

prompt

Generate <available_skills> XML prompt block for agent system prompts.

validator

Skill validation logic.