ubergoal

What problem does it solve?

Ubergoal removes ambiguity and missing evidence in substantial agentic coding work by making the goal owner accountable for tiering, routing, review-board coordination, acceptance proof, and post-run learning.

Core Features & Use Cases

• Goal-owning lifecycle wrapper: classifies risk tier, creates/binds a platform goal when available, and routes to the right subskills instead of monolith planning.
• Expectation/surprise gating: performs a user expectation vs surprise assessment early and re-checks during final readiness.
• Rigorous planning and bounded execution: routes planning to uberplan, coordinates specialist review-board lanes for Tier 2+, and enforces child-by-child operational outcome discipline.
• Failure-aware adaptation loop: stops and triggers RCA + scope correction when repeated or material unexpected test failures appear, rather than blindly patching.
• Evidence-first completion: routes final proof to uberaccept, prevents completion without policy-adherence checks, surprise delta reporting, and an Uber run receipt.
• Learning and skill evolution: hands off what actually worked (and what didn’t) to uberskillevolver.

Quick Start

Use the ubergoal skill when you have substantial coding or agentic-system work that needs safe tiering, specialist review-board lanes, rigorous planning, and evidence-backed final acceptance.