Demonstrates branching skill behavior on the current effort level using ${CLAUDE_EFFORT} (Claude Code v2.1.120+). Activates when discussing how to scope work across effort levels.
Activates when building or reviewing UI. Enforces WCAG 2.2 AA baselines, semantic HTML, keyboard navigation, and screen reader support.
Activates when designing HTTP/REST/GraphQL APIs, route handlers, or RPC interfaces. Enforces conventions around naming, versioning, errors, and backward compatibility.
Activates when investigating performance issues, profiling, or optimizing hot paths. Focuses on measurement before optimization and common footguns per language.
Activates when writing, modifying, or discussing tests. Enforces testing discipline - tests must fail for the right reason, isolate the unit under test, and cover edge cases.
Automatically triggered for code reviews, PR analysis, and quality checks. Use when reviewing changes, checking code quality, or preparing commits.
Automatically triggered for database migrations, schema changes, and data transformations. Use when working with database structure, migrations, or ORM models.