skill-authoring

Load offline Agent Skills defined by SKILL.md with required frontmatter fields.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/ririnto/sinon --skill skill-authoring-ririnto
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-authoring
Source: https://github.com/ririnto/sinon/tree/main/plugins/agent-capability-kit/skills/skill-authoring
Command: npx skills add https://github.com/ririnto/sinon --skill skill-authoring-ririnto

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Write or refactor one cross-platform Agent Skill so it is self-sufficient on activation, easy to trigger from its description, and usable offline without mandatory external tooling.

Core Features & Use Cases

  • Provides a flat, self-contained skill directory structure (SKILL.md plus optional references/, assets/, and scripts/ as needed) to make activation predictable.
  • Enforces mandatory frontmatter with at least name and description, enabling reliable discovery and offline loading.
  • Guides authors through a plan-validate-revise workflow, ensuring the ordinary path works from SKILL.md alone and optional assets deepen the context when blockers arise.

Quick Start

Write or refactor a single coherent Agent Skill by editing SKILL.md first and using references/assets/scripts only for named blockers or templates.

Frequently Asked Questions about skill-authoring

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

FAQPage Schema
How do I create an offline-friendly Agent Skill that works without external dependencies?

To create an offline-friendly Agent Skill, write a self-contained SKILL.md file with mandatory frontmatter fields (name, description) so it can be discovered and activated without external tooling. Use a flat directory structure with optional references, assets, and scripts to extend context.

What frontmatter fields are required for an Agent Skill to be discoverable?

Mandatory frontmatter for Agent Skill discovery requires at least the name and description fields. These fields enable reliable offline loading and predictable activation by allowing the agent to identify the skill and understand its trigger conditions.

How does the plan-validate-revise workflow help structure an Agent Skill project?

The plan-validate-revise workflow ensures the ordinary execution path works from SKILL.md alone before adding optional assets. Authors first validate the core logic, then deepen context with references, assets, or scripts only when named blockers arise, keeping activation predictable.

Can I use optional directories like references and scripts without breaking the ordinary skill activation path?

Yes, optional directories like references, assets, and scripts extend the skill without breaking the ordinary path. The core workflow executes from SKILL.md alone, while these optional folders provide deeper context or templates only when specific blockers arise during execution.

What is the best way to structure a self-contained Agent Skill directory for cross-platform use?

The best way to structure a cross-platform Agent Skill is using a flat, self-contained directory with SKILL.md at the root. Add optional references, assets, and scripts folders as needed to handle named blockers or templates without mandatory external tooling.

Why does my Agent Skill fail to trigger when loaded offline?

Agent Skills fail to trigger offline if mandatory frontmatter fields are missing or if the ordinary path relies on external tooling. Ensure SKILL.md contains at least the name and description fields and that the core workflow is self-sufficient.