R-PG-2

PowerGRYD — Layer Coverage

Layer Coverage at /reports/powergryd/coverage exposes the 17 RAIC source registers that feed the PowerGRYD pyramid, broken out by layer with row counts, last-updated stamps, and a top-3 status mix.

Canonical 17-source registry

The 17 sources are declared once in POWERGRYD_COVERAGE_SOURCES and consumed unchanged by this report and the Client Pack:

L1 Foundation: AIMS scope, leadership, objectives, governance policies.
L2 Data: ai_datasets + ai_system_datasets linkage.
L3 Models & Systems: ai_systems, use_case_register.
L4 Operations: controls, ai_system_monitoring_plans.
L5 Agentic: discovered_ai_apps with agent_type, agentic UCs (ai_tech_type overlap), AI Agent Acceptable Use policy.

Per-layer collapsible tables

Each of the 5 layers is rendered top-down (L5 → L1) as a collapsible section. Each row shows:

Source — register/table name.
Count — total rows for the org (via canonical safeCount).
Last updated — max(updated_at) with fallback to created_at.
Status mix — top-3 statuses (lowercased; nulls bucketed as __null__).

Direct deep-link: /reports/powergryd/coverage#L{n} auto-expands and scrolls that layer into view. This is the target of the Maturity report's "View coverage" buttons.

Errored-row warning

If any of the 17 source queries fail (network, RLS, permissions), the offending rows render with an amber warning chip and are excluded from layer rollups so a transient error never silently zeroes a score.

Export & audits

Gold-subtle "Export CSV" button produces RhindonCyber_PowerGRYD-LayerCoverage_*.csv (CRLF, MM/DD/YYYY HH:MM).

Audits: POWERGRYD_COVERAGE_VIEWED (60-min throttled), POWERGRYD_COVERAGE_EXPORTED. Tier/flag gating: Pro+Ent and feature_flags.powergryd.