addy-osmani

Production-grade engineering command suite for AI coding agents spanning the full development lifecycle — spec, plan, build, test, review, web performance, ship, and simplify.

Code review and quality enforcement workflow checking code style and patterns.

Code simplification workflow identifying opportunities to reduce cognitive complexity.

Incremental implementation workflow prioritizing execution of planned steps systematically.

Measurement-driven performance workflow: baseline with Lighthouse and RUM, identify bottlenecks via profiling, fix targeted issues (N+1 queries, render blocking, unoptimized images), verify against Core Web Vitals thresholds (LCP ≤2.5s, INP ≤200ms, CLS ≤0.1), and guard against regression with perf budgets.

Planning and task breakdown workflow decomposing features into manageable vertical slices.

Shipping and launch readiness checks for code deployment and integration.

Spec-driven development workflow enforcing specification generation before any coding starts.

Forces the AI agent to follow a strict red-green-refactor TDD workflow — explicitly blocking horizontal slicing (all tests first, then all code). Tests verify behaviour through public interfaces only, blocking code generation that skips the test step, and enforcing coverage thresholds before completing a task.