What problem does it solve? Contributors to the Summer skill library need a consistent way to verify that a skill meets structural standards and behaves as specified before committing it, without manual review of every rule. ## Core Features & Use Cases - Static lint mode: Checks frontmatter completeness, SKILL.md length limits, MCP-preferred patterns with file-edit fallbacks, collaborative protocol markers, template-id resolution, tests/spec.md presence, and relative link integrity. - Behavioral spec mode: Reads a skill's spec cases and reasons over whether the skill's instructions would satisfy each assertion, producing an annotated pass-rate report. - Library audit mode: Runs static checks across every skill registered in the plugin manifest and outputs a PASS/FAIL summary with a punch list. - Use Case: Before committing a new gameplay skill, run the static lint to catch a missing MCP example or broken reference link, then run the spec mode to confirm the skill's instructions satisfy its behavioral assertions. ## Quick Start Ask the agent to lint the skill at skills/<category>/<name> and report which structural checks pass or fail.
