# Independent Review Standard

Standard ID: `IRS-001-v0.1.0`
Status: governing reference (subordinate to the Category Intelligence Charter)
Adopted: 2026-07-26
Precedence: The [Category Intelligence Charter](CATEGORY-INTELLIGENCE-CHARTER.md) and the [Independence and Sponsorship Policy](INDEPENDENCE-AND-SPONSORSHIP-POLICY.md) govern this standard.

## Purpose

ETFCTA operates under a disclosed single-operator review model. That concentration (`RSK-008`) is the asset's largest institutional risk. This standard defines a **capability**, not a person: the governed process, records, and blocking gate that make an independent second review *mandatory and auditable* for the states where the stakes justify it — before any income or wider expansion depends on them.

No independent reviewer is appointed. The reviewer appointment status is `not_appointed` and stays so until a real, qualified person exists (see the [Reviewer Qualification Profile](REVIEWER-QUALIFICATION-PROFILE.md)). ETFCTA does not invent a reviewer to imply capacity it does not have.

## States that require an independent second review

An independent review is **mandatory** for any of the following review-triggering states:

1. `source_conflict` — a conflict between admitted primary sources (auto-detected when a decision's sufficiency is `conflict`).
2. `material_correction` — a correction that changes a published decision (auto-detected from a `correction` change event).
3. `ecs_amendment` — a material amendment to the ECS standard or a governed rule (auto-detected from a `rule_changed` change event).
4. `legal_or_regulatory_sensitivity` — a legally or regulatorily sensitive claim (identity, trademark, or dispute exposure).
5. `issuer_objection` — a formal objection from an issuer.
6. `paid_classification_report` — a paid report tied to a classification.
7. `commercial_relationship` — sponsorship or a commercial relationship with a party inside the category.

Triggers 1–3 are derived from the data and cannot be hidden by omission. Triggers 4–7 are declared on the decision or change via `review_triggers`. Declaring any trigger, or matching an auto-detected one, makes the state unpublishable until a valid approving review exists.

## The blocking gate

The `independent_review` check in `validation/registry-validator.mjs` (a required release check) blocks publication of any review-triggering state that lacks a valid, independent, approving review recorded in `governance/independent-review-register.json` and shaped by [`REVIEW-OUTCOME-SCHEMA.json`](REVIEW-OUTCOME-SCHEMA.json). A review authorizes publication only when it is `approved` or `approved_with_limitations`, the reviewer is independent of the classifier, no conflict is present, a scope and rationale are recorded, and a reviewed evidence trail is cited. The gate is enforced by `npm run verify:all` and `release:verify`.

## What a review must record

Every review record carries: reviewer identity, qualification basis, conflict presence and disclosure, scope reviewed, evidence reviewed, agreement, reasoning, limitations, date, and a signature or verifiable approval record — with one of the terminal outcomes `approved`, `approved_with_limitations`, `changes_required`, `not_resolved`, or `recused`. Only the first two authorize publication; the last three hold or block it. See the [Second Review Checklist](SECOND-REVIEW-CHECKLIST.md) and [Reviewer Decision Boundary](REVIEWER-DECISION-BOUNDARY.md).

## Boundary

This standard does not appoint a reviewer, change any current classification, or create any paid or sponsored surface. It makes the capability real in governance and code so that, the moment a qualified independent reviewer is appointed, triggered states can be reviewed and published under audit — and cannot be published otherwise.