import { type ConsensusRevision2Metadata } from "./statistics.js";
export type SkillBenchRunMode = "pilot" | "validated";
export type ReportDecisionState = "pass" | "fail" | "inconclusive";
type CellStatus = "complete" | "quality-failure" | "process-failure" | "infrastructure-failure" | "availability-failure" | "quota-failure" | "scorer-failure" | "incomplete" | "parity-invalid";
export interface ProofMatrixView {
    expected: string[];
    found: string[];
    done: string[];
    missed: string[];
    falsePositive: string[];
    incorrect: string[];
    proof: string[];
}
export type ReportConfidenceVerdict = "winner" | "tie" | "inconclusive";
export interface ReportConfidenceInput {
    verdict: ReportConfidenceVerdict;
    noWinnerReason: string | null;
    interval: {
        lower: number;
        mean: number;
        upper: number;
    } | null;
    metadata: ConsensusRevision2Metadata;
}
export interface ReportCellInput {
    id: string;
    taskId: string;
    arm: string;
    modelId: string;
    status: CellStatus;
    hardGatesPassed: boolean;
    qualityPassed: boolean;
    qualityScore: number;
    costUsd: number | null;
    latencyMs: number | null;
    samples: number;
    scenariosCovered: number;
    scenariosRequired: number;
    proofMatrix: ProofMatrixView;
    evidencePaths: string[];
    tokens: Record<string, unknown> & {
        completeness: string;
        provenance: string;
    };
}
export interface SkillBenchReportInput {
    schemaVersion: 1;
    runId: string;
    mode: SkillBenchRunMode;
    status: string;
    spec: {
        id: string;
        fingerprint: string;
        evaluationFingerprint: string;
        seed: string;
        rerunCommand: string;
    };
    skill: {
        id: string;
        fingerprint: string;
    };
    model: {
        id: string;
        fingerprint: string;
    };
    environment: {
        provider: string;
        fingerprint: string;
    };
    pricing: {
        source: string;
        url?: string | null;
        retrievedAt?: string | null;
        currency?: string | null;
        completeness: string;
    };
    budget: Record<string, number | string | null>;
    warnings: string[];
    confidence?: ReportConfidenceInput;
    cells: ReportCellInput[];
}
export interface ReportRankingCell {
    cellId: string;
    taskId: string;
    arm: string;
    modelId: string;
    qualityScore: number;
    costUsd: number | null;
    latencyMs: number | null;
}
export interface SkillBenchReportView {
    schemaVersion: 1;
    runId: string;
    mode: SkillBenchRunMode;
    decision: {
        mode: SkillBenchRunMode;
        state: ReportDecisionState;
        validated: boolean;
        cheapestPassing: ReportRankingCell | null;
        highestQuality: ReportRankingCell | null;
        matchedSkillUplift: number | null;
        coverage: {
            samples: number;
            scenarioCoverage: string;
        };
        decisiveMisses: string[];
        decisiveFalsePositives: string[];
        noWinnerReason: string | null;
        recommendedRoute: null | {
            skillId: string;
            modelId: string;
            objective: string;
        };
        taskChoices: Array<{
            taskId: string;
            state: ReportDecisionState;
            cheapestPassing: ReportRankingCell | null;
            highestQuality: ReportRankingCell | null;
        }>;
        confidence: {
            verdict: ReportConfidenceVerdict;
            noWinnerReason: string | null;
            familywiseAlpha: number;
            maxLooks: number;
            currentLook: number;
            comparisonFamilyId: string;
            comparisonCount: number;
            alphaPerComparisonLook: number;
            lowerQuantile: number;
            upperQuantile: number;
            frozenPairIds: string[];
            resamples: number;
            seed: string;
            sampleCount: number;
            coverage: number;
            interval: {
                lower: number;
                mean: number;
                upper: number;
            } | null;
        };
    };
    budget: Record<string, number | string | null>;
    tokenCostCompleteness: Array<{
        cellId: string;
        completeness: string;
        provenance: string;
        categories: string[];
    }>;
    proofMatrices: Array<{
        cellId: string;
        matrix: ProofMatrixView;
        evidencePaths: string[];
    }>;
    provenance: {
        manifestFingerprint: string;
        evaluationFingerprint: string;
        skillFingerprint: string;
        modelFingerprint: string;
        environmentFingerprint: string;
        seed: string;
        rerunId: string;
        rerunCommand: string;
        pricingSnapshot: SkillBenchReportInput["pricing"];
        warnings: string[];
    };
    recommendation: null | {
        evidencePath: string;
    };
    actions: {
        canApply: boolean;
    };
    rows: Array<ReportRankingCell & {
        status: CellStatus;
        statusText: string;
        hardGatesPassed: boolean;
        qualityPassed: boolean;
        tokens: ReportCellInput["tokens"];
    }>;
}
export declare function normalizeSkillBenchReport(input: SkillBenchReportInput): SkillBenchReportView;
export declare function renderSkillBenchReportJson(view: SkillBenchReportView): SkillBenchReportView;
export type RunOutcomeKind = "pilot" | "budget-stop" | "incomplete" | "salvaged" | "validated-pass" | "validated-fail" | "inconclusive";
export interface RunOutcomeView {
    kind: RunOutcomeKind;
    title: string;
    summary: string;
    nextActions: string[];
    technicalReason: string | null;
}
/** Plain-language run outcome for incomplete/budget-stop/pilot/pass paths. */
export declare function describeRunOutcome(view: SkillBenchReportView): RunOutcomeView;
export declare function renderSkillBenchReportHtml(view: SkillBenchReportView, options?: {
    browserOpen?: {
        attempted: boolean;
        ok: boolean;
        error?: string;
    };
}): string;
export {};