Skip to content
GridNinja

Deployment posture

Shadow Mode proves behavior before authority expands

Shadow Mode runs the capacity-acceptance path against live or replayed operating context without issuing authoritative setpoints. It records what GridNinja would have proposed, allowed, repaired, rejected, or marked no-proof, then compares those decisions with observed conditions and operator outcomes. The purpose is evidence accumulation and integration validation—not a disguised claim of autonomous operation.

Definition

What it is

Shadow Mode is a read-only deployment posture with explicit separation between observation, counterfactual decisions, and the systems that retain control authority.

Category boundary

What it is not

It is not a production control trial, a customer performance result, or permission to infer savings from counterfactual actions that were never executed.

Operational significance

Why operators care

Teams can test telemetry mappings, topology, constraints, refusal logic, and proof-pack quality while existing operating procedures remain authoritative. Evidence from this phase determines whether Advisory Mode is justified.

Mechanism

How the decision path works

  1. Step 01

    Ingest read-only telemetry and declared topology through scoped adapters.

  2. Step 02

    Generate timestamped counterfactual actions and runtime-assurance outcomes.

  3. Step 03

    Compare predicted constraints with observed operating states and operator records.

  4. Step 04

    Review false accepts, false rejects, no-proof cases, and data-quality failures before promotion.

Textual diagram equivalent

Decision flow

Read-only telemetry → counterfactual proposal → runtime-assurance outcome → comparison with observed operation → reviewed evidence → possible Advisory Mode gate.

Synthetic Shadow Mode review

Visibly scoped example

Synthetic illustrative scenario—not a customer or production result. A counterfactual workload shift is rejected after telemetry freshness crosses policy. No control command is sent; the stale-input refusal is retained for review.

Failure and no-proof cases

  • Shadow outputs can influence operators even when the authority boundary is poorly labeled.
  • Counterfactual benefits are reported as realized operational outcomes.
  • Mappings or timestamps are wrong, making comparisons appear precise but invalid.

Limitations

  • Shadow Mode cannot establish actuator behavior or closed-loop recovery that was never exercised.
  • Observed correlation does not establish that a counterfactual action caused an outcome.
  • Promotion requires named operational review and site-specific acceptance criteria.

Evidence and sources

Inspect the basis, not just the answer

Local evidence objects show the GridNinja proof contract. Primary sources provide external standards and risk-management context; they do not validate GridNinja performance.

Proof before autonomy

Test the evidence boundary against your capacity question

Start with a Capacity Audit or read-only Shadow Mode discussion. GridNinja does not require control authority to identify where a capacity claim remains unproven.