Task

This task is optional. Evaluate whether test-impact analysis (running only affected tests based on git diff) is worthwhile after Quick Win and Medium-term optimizations have already reduced PR time.

Background

Currently all 59 spec files run regardless of what files changed. A test-impact-map.json could map source paths to affected test files. Conservative default: full suite for unknown paths or >10 files changed.

This adds maintenance complexity and carries risk of false negatives (missed regressions). Nightly jobs always run the full suite as a safety net.

Phase 1: feasibility assessment (go/no-go)
Phase 2 (if go): implementation

Dependencies and Blockers

• Benefits from QW-4 (quarantine data helps distinguish false negatives from flakes)

Acceptance Criteria

• Feasibility document with go/no-go recommendation
• If go: mapping file, selector script, fallback to full suite, false negative tracking
• If no-go: rationale documented, task closed