skills-dev_skill

Overview

This skill helps developers design, initialize, and validate agent skills following the Agent Skills specification. It guides discovery, archetype selection, directory layout, frontmatter schema, and validation checks to ensure cross-platform compatibility. The content focuses on practical rules, file organization, and quality gates for publishable skills.

How this skill works

The skill inspects a skill's manifest and directory structure for required fields, naming conventions, and file presence. It runs a validation checklist against frontmatter syntax, description quality, structure limits, and platform-specific constraints, and produces a concise report with critical issues and warnings. It also recommends archetypes and progressive-disclosure patterns to keep runtime context small.

When to use it

• When creating a new agent skill or initializing a skill project
• When writing or editing a skill manifest and metadata
• When validating naming, frontmatter, or description quality before publishing
• When choosing an archetype for API wrappers, document processing, or dev workflows
• When enforcing cross-platform compatibility and platform-specific fields

Best practices

• Start with discovery: define problem, capabilities, triggers, and scope
• Pick an archetype (simple, api-wrapper, document-processor, dev-workflow, research-synthesizer) before scaffolding
• Keep the manifest concise: WHAT + WHEN + 3–5 trigger keywords, third-person voice
• Use a clear directory layout: manifest, scripts, references, assets; move details to references
• Follow strict naming rules: lowercase, numbers, hyphens, no leading/trailing hyphens, match directory
• Validate with a checklist: frontmatter syntax, description quality, file existence, and no placeholder markers

Example use cases

• Create a new dev-workflow skill that automates git branching and CI checks
• Author an api-wrapper skill for the GitHub or Stripe API with authentication notes in references
• Build a document-processor skill to extract tables from PDFs with extraction scripts in scripts/
• Validate a skill manifest and directory before committing to ensure parsing and loading at runtime
• Refactor a large skill by moving heavy docs and templates into references/ and assets/ to reduce context load