skill-creator_skill

This skill guides crafting Claude skills with concise, context-efficient workflows, references, and progressive disclosure for reliable agent customization.

58

GitHub Stars

2

Bundled Files

2 months ago

Catalog Refreshed

4 months 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 veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill ahonn/dotfiles --skill skill-creator

  • LICENSE.txt11.1 KB
  • SKILL.md17.6 KB

Overview

This skill guides you through creating, organizing, and packaging Claude skills. It explains the minimal required files, how to structure reusable resources, and a progressive disclosure pattern to keep context efficient. Use it to design skills that are concise, discoverable, and reliable for repeated agent usage.

How this skill works

The skill inspects the intended workflows and recommends what belongs in the SKILL.md frontmatter, what stays in the SKILL.md body, and what should be moved to scripts, references, or assets. It prescribes a creation process (understand, plan, initialize, edit, package, iterate) and patterns for when to load larger reference files so Claude only consumes necessary context. It also enforces constraints that keep skills token-efficient and focused.

When to use it

  • When building a new Claude skill from scratch
  • When updating or refactoring an existing skill to reduce context bloat
  • When deciding which resources to include as scripts, references, or assets
  • When packaging skills for reuse and deterministic execution
  • When defining triggers and concise frontmatter metadata

Best practices

  • Keep SKILL.md frontmatter clear and comprehensive—the name and description determine triggers
  • Put only essential workflows in SKILL.md; move large details to references and link them
  • Include scripts for deterministic or repeatedly rewritten code to save tokens and enable execution
  • Organize references by domain or variant so Claude only loads relevant files
  • Apply progressive disclosure: metadata → SKILL.md → bundled resources

Example use cases

  • Create a pdf-processing skill with scripts/rotate_pdf.py and references for advanced features
  • Refactor a large SKILL.md by moving API specs into references/api_docs.md and adding grep hints
  • Bundle a frontend template in assets/hello-world/ to speed up repeated scaffolding requests
  • Design a multi-provider deploy skill that links provider-specific references only when chosen
  • Initialize a new skill using init_skill.py to generate proper structure and placeholders

FAQ

Keep core procedural steps and selection guidance in SKILL.md; place large schemas, examples, and long docs in references and link to them.

When should I include scripts?

Include scripts when tasks require deterministic code or when the same code is repeatedly re-written—scripts save tokens and can be executed without loading into context.

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