skill-authoring

Guides authoring of Claude Code skill bundles with triggering descriptions and structured SKILL.md bodies.

Updated Jul 25, 2026
One-click install
npx skills add https://github.com/RorySullivan1/powerapp_taskmaster --skill skill-authoring-rorysullivan1
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-authoring
Source: https://github.com/RorySullivan1/powerapp_taskmaster/tree/main/.claude/skills/skill-authoring
Command: npx skills add https://github.com/RorySullivan1/powerapp_taskmaster --skill skill-authoring-rorysullivan1

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Writing a Claude Code skill that actually triggers and teaches well is hard: vague descriptions never fire, folder names mismatch the frontmatter, and bodies become reference dumps. This Skill provides the design method for creating, reviewing, and fixing skill definitions in .claude/skills/<name>/SKILL.md. ## Core Features & Use Cases - Layer and fold-first decisions: Decide whether a task warrants a skill versus an agent, command, context brief, or hook, and whether to extend an existing skill instead of creating a new one. - Trigger-optimized descriptions: Write descriptions that lead with identity, include concrete use-when scenarios, list trigger phrases, and draw boundary lines against sibling skills. - Structured body authoring: Apply the proven body shape of principles, method, worked examples, and watch-outs, plus a full authoring checklist and template. - Use Case: When a user asks "why isn't my skill triggering" or "write a skill for X", this Skill diagnoses description collisions, folder/name mismatches, and body bloat, then produces a corrected SKILL.md. ## Quick Start Ask the assistant to draft a new Claude Code skill for a recurring task, including the frontmatter description and body structure.

Frequently Asked Questions about skill-authoring

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

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

Write the description for triggering, not documentation: lead with a one-clause identity, add a concrete use-when scenario, list actual trigger phrases users say, and name implicit-trigger signals. End with explicit deferral clauses naming neighboring skills to prevent collisions.

When should I create a skill versus an agent or command?▼

Choose a skill for reusable expertise applied in the main conversation. Use an agent for isolated side tasks with summary output, a command for one-shot prompts, a context brief for reference facts, and a hook for deterministic lifecycle events.

Why isn't my skill triggering in Claude Code?▼

The description is usually too generic or collides with a sibling skill, causing both to under-fire. Also verify the folder name exactly equals the frontmatter name field, since a mismatch prevents the skill from resolving at all.

What fields are allowed in SKILL.md frontmatter?▼

Only name and description are allowed in skill frontmatter. Skills do not accept agent fields like tools, model, or permissions because a skill is advice loaded into the conversation, not an isolated runner.

Should I create a new skill or extend an existing one?▼

Default to extending an existing skill when there is real overlap, since skill proliferation degrades triggering for all skills. Only author a net-new skill after it passes significance gates for recurrence, non-obviousness, and generality.