/**
 * `prepareFinalReview()` — the 4-dimension business review primitive that gates
 * end-of-workflow human acceptance in the 10% human / 90% LLM workflow.
 *
 * Contract:
 *   - Reads the approved audit-goal JSON from
 *     `.peaks/_runtime/<sessionId>/audit-goal/<rid>.json`.
 *   - Calls an injected `LlmRunner` exactly once with a 4-dim review prompt.
 *   - Parses the LLM output as JSON and validates that the `dimensions`
 *     array covers ALL four required dimensions
 *     (functional-completeness, problem-resolution, no-new-bugs,
 *     existing-functionality-intact).
 *   - Throws `IncompleteFinalReviewError` if the JSON is malformed or any
 *     required dimension is missing. Callers MUST treat this as a gate
 *     failure (return to human for re-prompting) — autonomous work must
 *     never proceed on a partial review.
 */
export type DimensionKind = 'functional-completeness' | 'problem-resolution' | 'no-new-bugs' | 'existing-functionality-intact' | 'capability-consistency';
export type DimensionVerdict = 'pass' | 'fail' | 'inconclusive';
export type EvidenceKind = 'test-result' | 'test-coverage' | 'manual-spot-check' | 'pre-post-diff' | 'regression-suite' | 'ac-mapping';
export type DimensionConfidence = 'high' | 'medium' | 'low';
export interface EvidenceItem {
    readonly kind: EvidenceKind;
    readonly description: string;
    readonly artifact?: string;
    readonly link?: string;
}
export interface DimensionEvidence {
    readonly dimension: DimensionKind;
    readonly verdict: DimensionVerdict;
    readonly summary: string;
    readonly evidence: readonly EvidenceItem[];
    readonly confidence: DimensionConfidence;
}
export interface FinalReviewOutput {
    readonly rid: string;
    readonly generatedAt: string;
    readonly dimensions: readonly DimensionEvidence[];
    readonly overallSummary: string;
    readonly allPass: boolean;
    readonly needsAttention: readonly DimensionKind[];
}