What problem does it solve? Long-running coding agents often stop after easy work and leave quality debt—failing tests, lint errors, coverage gaps, outdated dependencies, and documentation drift—unaddressed. This Skill turns that debt into an explicit, verifiable work order with a deadline and a clock-out gate. ## Core Features & Use Cases - Quality Debt Composition: Scans the project's own tooling for lint, types, tests, flaky tests, coverage, dead code, TODO/FIXME debt, accessibility, localization, API contract drift, documentation drift, CI warnings, dependencies, and vulnerability advisories, then ranks and deduplicates findings. - Guided or Automatic Selection: Lets the owner pick quality areas (Guided) or lets Nightshift select every applicable high-value area that fits the available hours (Automatic), with review-first or run-direct launch modes. - Verified Shift Execution: Composes one ordered Hunt work order, arms a shift through the Start preflight, and holds unfinished work at a clock-out gate with receipts and token usage recorded. - Use Case: Ask the agent to spend four hours on quality debt; it runs the project's test and lint tools, ranks findings, composes a work order, and produces reviewable commits with per-item evidence. ## Quick Start Ask the agent to compose a quality shift for this project with a four-hour budget covering tests, lint, and dependency advisories.
