skill-architect

Creating a Claude Skill that activates precisely requires encoding domain expertise via shibboleths and anti-patterns. This meta-skill provides a 6-step creation process and subagent-aware design principles to ensure robust, efficient activation.

Progressive disclosure architecture in agent design enforces efficient loading of skill components. This meta-skill implements the architecture to ensure subagents load only necessary context, reducing overhead and improving response times.

Auditing an agent skill identifies common anti-patterns and verifies adherence to best practices. This meta-skill facilitates quality auditing by checking progressive disclosure compliance and evaluating domain expertise encoding.

The best way to structure agent skill creation is following a systematic 6-step workflow. This meta-skill guides users through defining requirements, encoding domain knowledge, implementing progressive disclosure, and validating activation behavior.

Python is required to build and audit Claude Agent Skills using this meta-skill. The framework includes Python scripts and reference components that automate the systematic design, creation, and quality auditing processes.

Skills load slowly or fail to activate due to poor progressive disclosure architecture and missing shibboleths. This meta-skill audits these constraints, ensuring efficient component loading and precise activation within subagent environments.