What problem does it solve? Writing effective agent skills requires knowing frontmatter conventions, progressive disclosure patterns, and defensive writing techniques, and evaluating third-party skills for security and quality is time-consuming without a structured process. ## Core Features & Use Cases - Skill Creation Workflow: Scaffold new skills with init_skill.py, write SKILL.md following proven patterns, validate with quick_validate.py, and package into distributable .skill files. - Third-Party Skill Evaluation: Run a 7-step evaluation covering source resolution, structural checks, security scanning, quality rubric scoring, context budget analysis, and report generation. - Defensive Writing Guidance: References cover testing methodology, rationalization tables, workflow patterns, and description refinement to make skills resilient against agent shortcut-taking. - Use Case: Before installing a skill found on GitHub, run the evaluation workflow to fetch the repo, scan it for security patterns like credential access or obfuscation, estimate its token cost, and produce an ADOPT/SKIP recommendation report. ## Quick Start Ask the agent to create a new skill called my-skill using the skill-building workflow, then validate it with the quick_validate script.
