cuit-looplisted

Close the UI feedback loop for an agentic coding model. Given a recorded session JSON (from the @cuit/recorder-extension Chrome extension, the @cuit/recorder npm module, or a third-party adapter), generate a Playwright/Vitest spec via @cuit/spec-gen, run it against the app, AUTO-DETECT which flow (baseline lock-in vs bug reproduction) the developer is in based on whether the spec passes on first run, then commit a GREEN regression gate. Use this whenever the user attaches a cuit-session JSON file or pastes session events.
speechlabinc/complex-ui-tester · ★ 1 · Testing & QA · score 77

Install: claude install-skill speechlabinc/complex-ui-tester

# cuit-loop — agentic UI feedback loop This skill turns a captured user session into a permanent CI regression gate. There are TWO flows the recorder supports, and the skill auto-detects which one the developer is in based on the result of the first spec run. ``` session.json │ ▼ @cuit/spec-gen │ ▼ run spec against current code │ ┌─────────────┴─────────────┐ │ │ PASS FAIL │ │ ▼ ▼ ┌──────────────┐ ┌──────────────┐ │ FLOW B │ │ FLOW A │ │ baseline │ │ bug repro │ │ lock-in │ │ │ └──────┬───────┘ └──────┬───────┘ │ │ │ ▼ │ read expected vs actual │ propose smallest code fix │ apply fix, re-run spec │ │ │ ┌─────────┴─────────┐ │ PASS FAIL │ │ │ │ ▼ ▼