autonomous-qa-orchestrationlisted

# Autonomous QA Orchestration This is the playbook for an open-ended "go test this" request. It sequences the other skills in this library rather than duplicating their content — pull each one in when you reach that stage. ## The pass 1. **Discover the application** — what is it, what stack, what environment (staging/prod/local), what's already known about it (existing memory/reports, `~/qa-agent/reports/`, prior findings). 2. **Understand architecture** — frontend/backend split, API shape, DB, third-party integrations, auth model. Skim rather than exhaustively read — enough to know where risk concentrates. 3. **Understand requirements/scope** — what changed recently (git log/diff, linked ticket/PRD) if this is feature-specific; if it's a general audit, use risk-based prioritization from [[qa-fundamentals]] to decide where to spend time. 4. **Explore the UI** — landmark tour per [[manual-testing]]'s exploratory testing method; use Claude in Chrome / Chrome DevTools MCP for live interaction. 5. **Explore the API** — capture real requests via network inspection, replay/probe per [[api-testing]]. 6. **Generate a test plan** — scoped to what was actually found in steps 1-5, using [[qa-fundamentals]]'s test plan structure; don't run a generic checklist disconnected from the actual app. 7. **Generate test cases** — via [[manual-testing]]'s design techniques and, where a schema/spec exists, [[qa-code-intelligence]]'s automatic generation approach. 8. **Execute** — manual where exp