This skill helps you run, debug, and extend the OpenClaw project's QA automation for channels like Telegram, WhatsApp, and Matrix. It reads scenario packs from the qa-lab extension, decides which lane to use (mock or live frontier), and runs full test suites with configurable model policies defaulting to gpt-5.4. You get structured JSON summaries and markdown reports in .artifacts, plus utilities for OTEL smoke tests, character evals across multiple LLMs, and credential management through 1Password and Convex. The workflow is opinionated: it won't use gpt-5.4-pro or mini unless you ask, and it expects you to rerun the full lane if anything fails. Useful when you're shipping transport features or validating end-to-end channel contracts before release.
npx skills add https://github.com/steipete/clawdis --skill openclaw-qa-testing