fix: align latest assessments with Platform findings

[dweinstein]

Summary

• default list_assessments to findings-capable NowSecure Platform assessments
• add platform, store_monitor, external, and all track selection using upstream pre-pagination filters
• classify assessment sources explicitly and mark external/store-monitor rows unavailable to get_assessment_findings
• update MCP guidance, schema snapshot, README, and regression coverage

Why

The newest mixed-history rows can be store-monitor assessments with no Platform findings. That makes list_assessments disagree with list_apps.assessment_ref and leads agents into a guardrail error when fetching findings.

Verification

• go test ./...
• go vet ./...
• git diff --check