SPEC alignment audit and gap fix — Phase 5 skills, infra, docs

[dtsong]

Summary

• 3 new skills: data-testing, data-governance, data-observability (Phase 5 of roadmap)
• Infrastructure gaps fixed: pipeline/README.md, ENGAGEMENT.yaml template, security-tier-model expanded with Tiers 0-4
• Cross-references wired: handoff links added to tsfm-forecast, dbt-transforms, client-delivery, data-pipelines
• install.sh updated: 13 skills, fixed data-platform-engineer role (was missing 7 skills), added data-testing to data-consultant
• CATALOG.md and README.md updated: new category, sizing table, badge (9→13), roadmap Phase 5 marked complete

New Skills Detail

Skill	Files	Lines	References
data-testing	7	850	dbt-testing-strategy, pipeline-validation, sql-assertion-patterns, test-as-deliverable
data-governance	8	1,110	data-catalog, lineage-patterns, classification-policies, access-control, compliance-frameworks
data-observability	8	1,160	freshness-monitoring, volume-anomaly-detection, schema-change-detection, alerting-patterns, incident-response

All three skills include trigger and navigation eval cases.

Test plan

• ./install.sh --list shows 13 available skills with correct role mappings
• pre-commit run --all-files passes (all hooks green)
• python3 pipeline/hooks/check_references.py — no broken references
• python3 pipeline/hooks/check_context_load.py — new skills under 5,500 token ceiling
• grep -r "data-governance-skill\|data-observability-skill\|data-testing-skill" --include="*.md" . — zero results (no stale -skill suffix)
• Verify cross-reference paths resolve in tsfm-forecast, dbt-transforms, client-delivery, data-pipelines

🤖 Generated with Claude Code