agentic-ux

The reference for designers shipping AI agents — design × AI, inclusive by default.

Seven interface patterns for AI products people can trust, each with a working demo: sources, confidence, editable output, undo, why we ask, boundaries, review before acting.

Agentic UX & WebMCP Auditor CLI by EchoRank

An intelligent, zero-trust agentic UX layer for Android that intercept and filters high-friction notification storms using on-device hardware-accelerated AI.

Mission control for autonomous agents. One event stream, five surfaces.

🛠️ Decoupled agentic UX/UI framework — a pure-Python foreman orchestrating Goose (research/scrape) + Aider (codegen) through a 4-phase engine. Two modes: niche creation (--target) and live site clone+modify (--url --modify). No LangChain, no CrewAI — stdlib + litellm only.

A concept for an agent-driven banking use-case. Describe what should happen with your money once, in plain language. An agent reads your account, builds the routine, and shows every assumption it made before anything runs.

UX principles for AI agents that act for a person: autonomy earned tier by tier on evals, the write contract with the system of record, dense operator screens. 4 skills, 2 diagrams.