Give your AI agent a product judgment layer before the next build.
UX3 Product Design Harness helps product leaders, founders and designers review a feature, experiment, launch or feedback signal with their existing AI agent. Review User Flow, Evidence Flow and Business Flow; return one continue, verify or stop_reframe verdict, with the weakest flow, evidence gaps and a bounded next action. The complete MIT-licensed skill includes prompts, schemas, rules, templates, validators and worked examples. Validators check structure and consistency, not whether supplied evidence is true. Humans remain accountable for decisions. Free to install from GitHub; requires an AI agent, plus Python 3 and jsonschema for validation. Connect via LinkedIn to discuss team adoption.
-
Three connected reviews: users, evidence and business viability
-
One verdict with explicit execution boundaries and stop conditions
-
Quick Gate, Standard Gate and UX3 Council review modes
-
Complete prompts, schemas, rules, templates and validators
-
Free MIT license; source and worked examples included
-
Review a proposed feature before committing engineering time
-
Design a bounded experiment and identify disconfirming evidence
-
Assess launch readiness across user value and business viability
-
Turn user feedback into a clear product decision
-
Document an accountable handoff from product review to execution