authoring-skills_skill

This skill helps you create, structure, and maintain agent skills under .agents/skills by guiding frontmatter, content, and naming conventions.
  • JavaScript
  • Official

138.1k

GitHub Stars

1

Bundled Files

3 weeks ago

Catalog Refreshed

1 month ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstart where the catalogue uses aiagentskills.

npx veilstart add skill vercel/next.js --skill authoring-skills

  • SKILL.md4.0 KB

Overview

This skill explains how to create and maintain agent skills in the .agents/skills/ directory. It helps you decide when to split content out of the always-loaded agent file into on-demand skills and how to structure metadata, naming, and supplementary files for reliable auto-loading and discoverability.

How this skill works

It inspects and prescribes the required frontmatter fields and values the agent runtime recognizes and shows how description text drives auto-activation. It guides how to organize content into a short hub file plus optional detail pages, what belongs in the always-loaded agent file versus an on-demand skill, and how to reference skills from the main agent surface.

When to use it

  • When content is too detailed for the always-loaded agent surface (multi-step workflows or code templates).
  • When information applies only to specific tasks and should load on demand.
  • When you need a self-contained troubleshooting or diagnostic procedure.
  • When adding step-by-step verification commands and examples for a pattern.
  • When you want to keep short guardrails in the always-loaded file but expand details separately.

Best practices

  • Make the description the primary matching surface: state what it covers, when to use it, key file names, and matching keywords.
  • Lead with 'Use this skill when...' and include concrete, numbered steps and verification commands.
  • Use only the supported frontmatter fields and avoid unknown fields to ensure runtime compatibility.
  • Name skills short and topic-scoped with hyphens for multi-word names (no repo prefixes).
  • Add a one-line summary in the always-loaded agent file that references the skill by name.

Example use cases

  • Documenting an end-to-end feature flag change with file list, code snippets, and verification commands.
  • Providing a DCE pattern with full code examples, edge cases, and CI checks separate from guardrails.
  • Creating a triage hub that links to workflows, repro steps, and prioritization guidance.
  • Packaging a local reproduction guide for flaky CI that includes environment setup and debug commands.

FAQ

Only use the recognized fields listed by the runtime: name, description, argument-hint, user-invocable, disable-model-invocation, allowed-tools, model, context, and agent.

How do I make a skill auto-load reliably?

Write a specific description that includes the topic, trigger scenarios, key filenames, and likely keywords so the matcher can trigger the skill when relevant.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational