# GridNinja Sample Proof Pack

This illustrative sample previews the operator-facing artifacts behind Proof Before Autonomy. It is not validated site evidence.

## Deployment boundary
- Mode: read-only Shadow Mode
- Write credentials: none
- Command VLAN requirement: none
- Authority: operator policy, deterministic checks, and runtime assurance remain the gate

## Proof thesis
Every accepted MW must point to a proof row. If evidence is stale, incomplete, or outside policy, the correct answer is no-proof.

## Included artifacts
- AI Data Center Load Passport
- Capacity Waterfall
- RTA Allow / Repair / Reject / No-Proof Trace
- Reserve-Floor Report
- No-Proof Gap Register
- Utility Evidence Packet

## Evidence classes
- Technical data: telemetry completeness, constraint maps, replay traces, RTA outcomes
- Operational data: operator review time, no-proof gaps, policy exceptions, workload/cooling constraints
- Commercial data: proof-adjusted MW, accepted-headroom ledger, design-partner criteria
- Impact data: grid-readiness posture, dispatch envelope, local reliability support potential

## Sample log
- 14:05:12: REPAIR Cluster cap reduced 4%
- Reason: PCC margin projected below threshold
- Constraint: Feeder thermal envelope
- Margin: +3.2%
- SLA impact: None predicted
- Proof posture: Accepted after repair, not autonomous approval

## No-proof posture
- Missing telemetry is assigned to an owner before capacity is counted.
- Unknown topology is a proof gap, not hidden headroom.
- Reserve-floor policy is declared before bridge power or storage is considered flexible.

## Review checklist
- Binding constraint, remaining margin, and reason code are visible.
- The action bundle is written in operator language, not model jargon.
- No-proof gaps are treated as safety outcomes, not hidden uncertainty.
- Replay, rollback, and audit evidence are readable without a dashboard.

## Public / private boundary
- Public: artifact types, methodology summary, non-confidential findings, and source notes.
- Private: site topology, credentials, customer data, security-sensitive telemetry, and exact operating limits.

## Next step
Request a Capacity Audit to generate a site-specific proof pack and Shadow Mode evidence baseline.