What problem does it solve?
Kova developers need a repeatable way to run, watch, and debug QA-lab and qa-channel end-to-end scenarios, including live OpenAI lanes and matrix coverage, without losing time to setup, failures, or missing artifacts.
Core Features & Use Cases
- QA Suite Execution for qa-lab/qa-channel: Guides selecting the correct lane (mock/dev vs live-frontier) and produces deterministic artifact outputs for review.
- Live Run Monitoring & Failure Recovery: Specifies where to find the suite summary/report, how to locate the QA listen port for watching live UI, and what to do when a scenario fails.
- Specialized Validation Flows: Covers OTEL smoke validation, Matrix lane sharding for transport/media/e2ee coverage, and character evaluation to judge persona/style consistency across models.
- Credential & Environment Handling: Documents how to source QA secrets from 1Password and how to map provider modes and Telegram/Docker/Convex requirements correctly.
Quick Start
Ask the AI agent to run the live OpenAI QA suite using gpt-5.4 fast mode and output artifacts to .artifacts/qa-e2e for your chosen tag.