skill-creator

Creating an AI skill involves defining goals, outcomes, and reusable patterns in a structured format with frontmatter constraints (hyphen-case name, description under 500 characters), organized sections (Overview, When to Use, Core Pattern, Common Pitfalls), and discovery keywords. Use the init_skill.py script to scaffold a new skill, then populate scripts, references, and assets for consistency and agent discoverability.

Organize skills with lightweight references, structured sections, and discovery keywords in a centralized repository. Skills require frontmatter metadata, optional supporting files, and minimal embeddable examples and tests. This standardization ensures agents can locate and reuse skills efficiently across projects.

Create a custom skill when you need to extend Claude's capabilities with company-specific knowledge, complex multi-step workflows, or new tool integrations. Skills are designed for scenarios where you want reusable, discoverable, and well-structured patterns that can be applied consistently across multiple projects.

Enforce frontmatter constraints: skill names must be in hyphen-case, descriptions limited to 500 characters, and no angle brackets allowed. Follow structured patterns with clear outcomes, provide minimal examples and tests, and organize scripts, references, and assets to optimize context management and AI efficiency.

Yes. Skills are designed as reusable units with triggers and standardized patterns that work across projects. The structured format, discovery keywords, and lightweight references enable agents to locate and apply skills consistently in different contexts and workflows.

Include discovery keywords in your skill's metadata, use the standardized frontmatter format, and organize content in structured sections. The init_skill.py script generates this scaffolding automatically, and proper naming conventions (hyphen-case) and clear descriptions make skills easily searchable and indexable by agents.