Playwright
Trigger Boundary
- •Use when real browser interaction is required to validate behavior.
- •Do not use for unit or API-only verification scopes.
- •Do not use for desktop app capture where no browser is involved.
Goal
Produce deterministic browser-run evidence for functional and UI flow validation.
Inputs
- •Target URLs and critical user journey definition
- •Test credentials or fixtures (provided via environment variables or secret store)
- •Required artifacts (
snapshot,screenshot,trace)
Outputs
- •Reproducible command sequence for target flow
- •Artifact bundle for key checkpoints and failures
- •Findings list with exact reproduction references
Workflow
- •Confirm Playwright CLI runtime availability and browser readiness.
- •Prepare test credentials from environment variables; avoid inline plaintext secrets.
- •Open target page and snapshot before interacting.
- •Execute flow step-by-step; refresh references after DOM-changing actions.
- •Capture artifacts at decision points and failure boundaries.
- •Report observed behavior with command-level traceability.
Quality Gates
- •Every action maps to a fresh, valid UI reference.
- •Flow can be replayed from the documented command sequence.
- •Artifacts cover both success path and failing branch when present.
- •Conclusions are based on observed runs, not inferred behavior.
- •Credentials are never hardcoded in commands, scripts, logs, or screenshots.
Failure Handling
- •Stop when automation prerequisites (runtime, auth, env) are unmet.
- •Escalate when deterministic replay is impossible due to external instability.
- •Stop when only plaintext credentials are available and secure injection is not possible.