Documenting deep dives into performance optimization, AI prompt engineering, and modern web architecture.
Claude Project Autonomy gives Claude Code persistent, project-scoped development permissions so it can build, test, debug, install dependencies, and use the web without repeatedly stopping for routine approvals.
A practical setup for making Claude Code understand a large codebase through a local knowledge graph before it starts blindly searching files.
A practical system for reducing sycophancy in large language models during serious problem-solving work — a protocol that makes a model evaluate corrections on evidence, not agreement.
An open, repository-local product-design skill that teaches coding agents how to make evidence-driven decisions for AI products, research interfaces, RAG, model evaluations, and automation workflows.
A surgical approach to achieving 100/100 Lighthouse scores and perfect Core Web Vitals using an advanced AI optimization engine.