skill-forge

Create distributable Claude Code skills with structured SKILL.md and .skill packaging.

Updated Apr 15, 2026
One-click install
npx skills add https://github.com/ToanPV90/dotfiles --skill skill-forge-toanpv90
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-forge
Source: https://github.com/ToanPV90/dotfiles/tree/main/agents/.agents/skills/skill-forge
Command: npx skills add https://github.com/ToanPV90/dotfiles --skill skill-forge-toanpv90

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Skill Forge prevents Claude Code skills from becoming unstructured, inconsistent markdown dumps by giving you a production-grade, workflow-driven template for writing SKILL.md that triggers reliably and executes safely.

Core Features & Use Cases

  • Keyword-rich description engineering: builds trigger coverage so the skill activates when users phrase requests naturally for the capability.
  • Token-efficient skill architecture: keeps SKILL.md lean (<500 lines) via progressive loading from references.
  • Deterministic workflow design with guardrails: uses an Iron Law, a trackable checklist, confirmation gates, and pre-delivery quality checks to reduce shortcuts and execution mistakes.
  • Resource planning patterns: separates deterministic operations into scripts, deep knowledge into references, and output-only items into assets.
  • Packaging support for distribution: validates and packages the skill into a distributable .skill file.

Quick Start

Run the skill at /skill-forge and follow the guided workflow to design and package your next distributable skill.

Frequently Asked Questions about skill-forge

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

FAQPage Schema
How do I create a Claude Code skill that triggers reliably?

To create reliable Claude Code skills, use a workflow-driven SKILL.md template with keyword-rich descriptions, an Iron Law constraint, and confirmation gates. This enforces trigger coverage and execution safety, preventing unstructured markdown dumps.

What is progressive loading in skill design?

Progressive loading in skill design keeps SKILL.md lean by loading deep knowledge on-demand from external references. This token-efficient architecture keeps the main file under 500 lines, reducing context window consumption during execution.

How do I package a Claude Code skill for distribution?

Packaging a Claude Code skill involves separating deterministic operations into scripts, deep knowledge into references, and validating the folder structure before archiving it into a distributable .skill file.

How do I enforce quality and safety when writing Claude Code skills?

You enforce quality and safety in Claude Code skills by applying deterministic workflow guardrails, including a trackable execution checklist, confirmation gates, and pre-delivery quality validation checks to reduce shortcuts and mistakes.

Do I need YAML frontmatter for a Claude Code skill?

Yes, mandatory YAML frontmatter with name and description fields is required for Claude Code skills. This metadata supports prompt engineering by ensuring the skill activates naturally based on user requests.

What is the best way to structure a large Claude Code skill?

The best way to structure a large Claude Code skill is resource planning: separating deterministic operations into scripts, deep knowledge into references, and output-only items into assets. This maintains a lean SKILL.md core file.