unlazy

What problem does it solve? Agents often report work as done when it is half-finished, silently reduce scope, or claim unmeasured results. This Skill makes incomplete work visible by requiring acceptance gates written before execution, runnable checks with explicit approval, and independent re-verification of evidence before any completion report. ## Core Features & Use Cases - Acceptance Gate Ledgers: Write one observable outcome per gate with CHECK and EXPECT clauses, then run them through a zero-dependency Node checker that requires both exit code 0 and output matching. - Depth Tree Orchestration: Decompose large builds or audits into leaf and branch ledgers with ownership leases, scoped pipelines, and atomic dispatch waves for parallel agent work. - Fail-Closed Verification: Re-verify returned work with --reverify, treat abandoned gates as honest handoffs rather than success, and audit the final report against the original request. - Use Case: Before a multi-part refactor, create GATES.md from the template, approve the reviewed checks, dispatch leaf work in waves, and report only measured met, unmet, and abandoned gate counts. ## Quick Start Ask the agent to use the unlazy skill to write acceptance gates for your task, then run the gate checker with explicit approval before starting implementation.