7 principles for unicorn-grade product engineering. TDD-validated Claude Code SKILL extracted from 56 articles analyzing Claude Code internals.

Architecture Decision Quality Benchmark — the first public benchmark for evaluating AI on architecture-level engineering decisions

Policy-aware code execution orchestration for AI-native applications.

Nebutra AI runtime for multi-agent orchestration and Vercel AI SDK helpers.

Nebutra design-system UI primitives, layouts, and token-integrated components.

Nebutra icon exports and generated Geist-compatible icon components.

Model Context Protocol primitives for Nebutra agent and tool surfaces.

Structured logging and OpenTelemetry instrumentation for Nebutra services.

Outbound webhook contracts and delivery primitives for Nebutra services.

Registry and loader primitives for agent skills, tools, and local capability packs.

Self-hosted OSS font registry for Nebutra themes and DESIGN.md imports.

Runtime CSS variables and theme provider primitives for Nebutra products.

Sandbox runtime contracts for governed AI tool execution.

Cache provider contracts and runtime helpers for Nebutra applications.

Typed error definitions and platform error contracts for Nebutra packages.

Provider resolution primitive for explicit, environment, detection-chain, and fallback selection.

Transactional email contracts and provider integration helpers.