skill-authoring_skill

Overview

This skill guides authors through creating and maintaining user-facing agent skills for B2C CLI and tooling. It focuses on clear discovery metadata, folder organization, progressive disclosure, and writing concise, actionable instructions so skills trigger correctly and remain maintainable.

How this skill works

The skill inspects and enforces a consistent skill folder layout, naming conventions, and a required frontmatter block that summarizes what the skill does and when to use it. It reviews instruction structure, examples, troubleshooting, and optional resources like scripts, references, and assets. It also provides a validation checklist and iteration guidance to fix under- or over-triggering and execution issues.

When to use it

• Creating a new user-facing agent skill for CLI or platform workflows
• Writing or revising a skill’s discovery metadata and short description
• Organizing skill directories, scripts, reference docs, and static assets
• Improving a skill’s trigger behavior (undertriggering or overtriggering)
• Preparing skills for production testing and monitoring after deployment

Best practices

• Define 2–3 concrete use cases up front with triggers, steps, and expected results
• Use kebab-case for folder and skill names and keep file names exact and consistent
• Make the description front-and-center: state what it does, when to use it, and include task-oriented trigger phrases
• Limit the main instructions length and move heavy detail to targeted reference files
• Provide concise, actionable examples and troubleshooting entries for common errors
• Include small, self-contained scripts with robust error messages when automation is required

Example use cases

• Author a skill that helps troubleshoot B2C CLI authentication errors: triggers include "auth failed" or "connection error"
• Create a lookup skill for B2C Commerce API tasks: triggers like "how do I generate a URL" or "query products"
• Build a deploy checklist skill with step-by-step commands and verification checks
• Refactor an existing skill to reduce over-triggering by adding negative triggers and clearer scope
• Add reference files for schemas and API patterns to keep the main instructions short