What problem does it solve? Teams repeatedly encode the same AI usage patterns and project conventions, but without a consistent structure, skills end up as bloated prose documents that LLMs cannot reliably activate or follow. ## Core Features & Use Cases - Standardized Skill Scaffolding: Generates SKILL.md files with required frontmatter (name, description, license, metadata) and ordered sections: Activation Contract, Hard Rules, Decision Gates, Execution Steps, Output Contract, References. - Style Guide Enforcement: Applies the repo's normative skill style guide (docs/skill-style-guide.md) with a bundled fallback copy and compact inline rules when guides are unavailable. - Token Budget Control: Enforces a 180–450 token body target with a 1000-token hard maximum, moving supporting material into assets/ or references/ directories. - Use Case: When your team notices agents repeatedly mishandling Jira ticket creation, use this Skill to scaffold a jira-task skill with trigger-word frontmatter, decision tables, and registration in AGENTS.md. ## Quick Start Ask the agent to create a new skill for a repeated workflow, such as "create a skill for writing database migration scripts following our team conventions."
