Get configuration and recommendations for a specific stage

Get the recommended prompt structure for a specific stage

Get recommended model (Sonnet for analysis, Haiku for execution)

Save a clean output at the end of each stage for next stage input

Load clean output from previous stage (not full chat)

Check if context is clean and not bloated

Create clean handoff to start new session

Save test cases for future reuse on bug fixes/regressions

Find existing test cases for reuse

Check if test cases are ready for execution

Validate output before moving to next stage

Create a new test case with title, description, steps, and expected result

Record the result of a test execution

Report a new defect/bug found during testing

Generate comprehensive QA report with all metrics and statistics