agent-skills

Package agent capabilities with SKILL.md metadata for discovery and on-demand loading.

Updated Aug 23, 2026

Search Tags:#progressive-disclosure #ai-agents #agent-skills #skill-structure #open-standard #markdown-instructions

Ready to run

npx skills add https://github.com/quantmind-br/skills --skill agent-skills-quantmind-br

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-skills
Source: https://github.com/quantmind-br/skills/tree/main/agent-skills
Command: npx skills add https://github.com/quantmind-br/skills --skill agent-skills-quantmind-br

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agent Skills provide a portable, open-standard format for packaging and discovering reusable agent capabilities that can be loaded on demand by compatible agents.

Core Features & Use Cases

Portable, version-controlled skill packages that can be shared across AI agents and tools
Progressive disclosure: lightweight startup metadata with full instructions loaded at activation
Cross-platform interoperability: supports scripts, references, and assets to extend capabilities

Quick Start

Install the Agent Skills folder with a SKILL.md at the root and begin by discovering and loading its metadata.

Frequently Asked Questions about agent-skills

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

What is an open standard for packaging AI agent capabilities?▼

An open standard for packaging AI agent capabilities provides a portable, version-controlled format that standardizes how reusable skills are discovered and loaded on demand by compatible agents.

How do I package reusable AI agent skills for on-demand loading?▼

To package reusable AI agent skills, create a folder with a SKILL.md file at the root containing YAML frontmatter for validation, and optionally include scripts, references, and assets to extend capabilities.

Does progressive disclosure help with AI agent startup performance?▼

Progressive disclosure improves AI agent startup performance by loading only lightweight metadata initially, deferring full skill instructions until the specific capability is activated on demand.

Can I use version-controlled skill packages across different AI agent platforms?▼

Yes, version-controlled skill packages support cross-platform interoperability, allowing portable capabilities with scripts, references, and assets to be shared across different AI agents and tools.

What YAML frontmatter validation is required for an agent skill?▼

YAML frontmatter validation for an agent skill requires mandatory name and description fields, while optional fields support scripts, references, and assets for progressive-disclosure loading workflows.