add Browser QA Agent open questions

Author: BAKUGOS1

examples/browser-qa-agent/01-open-questions.md

@@ -0,0 +1,40 @@
+# Open Questions
+
+## Product
+
+- Should v1 run as a CLI only or include a web dashboard?
+- Should reports be Markdown-only, JSON-only, spreadsheet-compatible, or all three?
+- Should screenshots be embedded in report output or linked as local files?
+- Should the agent classify issues as bug, UX issue, feature gap, performance issue, or blocker?
+- Should the agent support test plans written in Markdown?
+
+## Browser Workflow
+
+- Which workflows are approved for default testing?
+- Should the agent auto-discover navigation or follow an explicit route list?
+- Should it use visible text selectors, test IDs, or a hybrid selector strategy?
+- Should mobile viewport testing be part of MVP?
+- Should console/network monitoring be enabled by default?
+
+## Safety
+
+- Which actions are considered destructive for the target app?
+- Should archive/restore be allowed by default while permanent delete is blocked?
+- Should the agent require confirmation before submitting forms?
+- What should happen if test account sign-in fails?
+- Should the agent mask email, phone, or customer data in reports?
+
+## Technical
+
+- Where should screenshots be stored?
+- Should traces be captured for every run or only on failure?
+- Should output include machine-readable bug IDs?
+- Should reports be deterministic for CI usage?
+- What timeout and retry policy should be used per step?
+
+## Evaluation
+
+- What fixture apps should be used for regression testing?
+- How do we score whether the agent found the correct bug?
+- How do we test that the agent avoids destructive actions?
+- How do we compare agent reports across versions?