What problem does it solve? Writing an agent skill that actually gets discovered and triggered is hard: descriptions without triggers never load, folder names mismatch frontmatter, and bloated SKILL.md files blow the activation token budget. This Skill provides a step-by-step authoring procedure so your skills conform to the agentskills.io specification and load correctly. ## Core Features & Use Cases - Skill Scaffolding: Create a new skill from scratch with the correct folder layout (SKILL.md plus optional scripts/, references/, and assets/ directories) and a minimal skeleton template. - Frontmatter & Naming Rules: Enforces naming constraints (lowercase, hyphens, max 64 chars, folder name must match the name field) and shows how to write descriptions that answer what the skill does and when to load it. - Progressive Disclosure Design: Explains the three loading stages (discovery, activation, execution) and how to split content into focused reference files to stay within token budgets. - Validation & Auditing: Provides a manual checklist and the skills-ref CLI command to validate a skill, plus a catalog of common mistakes like hidden script dependencies and vague instructions. - Use Case: You built a pdf-processing skill but the agent never triggers it. Load this Skill to audit the description for missing trigger phrases, verify the folder name matches the frontmatter, and split an oversized SKILL.md into references/. ## Quick Start Ask the agent to scaffold a new agent skill named 'code-review' with proper frontmatter, folder layout, and a trigger-rich description following the agentskills.io specification.
