import type { ExecutionTokenUsage } from './execution-usage.js'; export declare const COMPARATIVE_BENCHMARK_SCHEMA: 1; export type BenchmarkSystemKind = 'VANILLA_OPENCODE' | 'OPENCODE_HI_BASELINE' | 'OPENCODE_HI_CURRENT' | 'EXTERNAL_BASELINE'; export type BenchmarkEpisodeKind = 'DETERMINISTIC_FIXTURE' | 'REAL_HOST_EPISODE' | 'POLICY_ABLATION'; export type BenchmarkScenarioClass = 'trivial-localized-work' | 'failing-test-fix' | 'independent-parallel-work' | 'dependency-fan-in' | 'mutable-surface-conflict' | 'verification-route-unavailable' | 'independent-review-required' | 'misleading-done' | 'mutation-after-verification' | 'provider-child-failure' | 'restart-stale-callback' | 'authority-ambiguous-replay' | 'context-heavy-investigation' | 'plugin-config-coexistence' | 'production-commit-task'; export type BenchmarkCheckStatus = 'PASS' | 'FAIL' | 'BLOCKED' | 'NOT_RUN'; export type BenchmarkResultClassification = 'VERIFIED_SUCCESS' | 'VERIFIED_FAILURE' | 'BLOCKED_ENVIRONMENT' | 'BLOCKED_AUTHORITY' | 'TIMEOUT' | 'INVALID_RECEIPT'; export interface BenchmarkSystemIdentity { kind: BenchmarkSystemKind; label: string; opencode_version: string; opencode_commit?: string; hi_commit?: string; external_version?: string; config_sha256: string; } export interface BenchmarkTaskIdentity { task_id: string; scenario_class: BenchmarkScenarioClass; fixture_sha256: string; repo?: string; from_commit?: string; to_commit?: string; } export interface BenchmarkModelIdentity { requested?: string; effective?: string; provider_effective?: string; } export interface BenchmarkDeterministicCheck { id: string; status: BenchmarkCheckStatus; exit_code?: number; evidence_refs: string[]; detail?: string; } export interface BenchmarkEvidenceSummary { required: number; satisfied: number; fresh: number; stale: number; wrong_task_accepted: number; wrong_attempt_accepted: number; false_completion: number; } export interface BenchmarkControlPlaneMetrics { duplicate_dispatch_count: number; stale_callback_accept_count: number; ambiguous_side_effect_replay_count: number; deadlock_or_stall_count: number; orphan_or_cleanup_failure_count: number; workers_spawned: number; retries: number; replans: number; tool_calls: number; model_calls: number; polling_calls: number; peak_concurrent_workers: number; context_bytes_to_children: number; mechanically_identified_redundant_actions: number; } export interface BenchmarkExactUsage { tokens: ExecutionTokenUsage; coverage: 'COMPLETE_STEP_TOTAL' | 'PARTIAL_MESSAGE_REPORTED'; source: 'OPENCODE_STEP_FINISH' | 'OPENCODE_ASSISTANT_MESSAGE' | 'PROVIDER_USAGE'; } export interface BenchmarkEstimatedUsage { tokens?: number; cost_usd?: number; method: string; } export interface BenchmarkEconomics { wall_time_ms: number; exact_usage?: BenchmarkExactUsage; estimated_usage?: BenchmarkEstimatedUsage; provider_billed_cost_usd?: number; opencode_derived_cost_usd?: number; } export interface BenchmarkFailureInjection { id: string; kind: string; applied: boolean; observed: boolean; } export interface BenchmarkArtifacts { diff_sha256?: string; acceptance_log_sha256?: string; receipt_inputs_sha256: string; } export interface ComparativeBenchmarkReceipt { schema: typeof COMPARATIVE_BENCHMARK_SCHEMA; episode_kind: BenchmarkEpisodeKind; claim_boundary: string; episode_id: string; repetition: number; system: BenchmarkSystemIdentity; task: BenchmarkTaskIdentity; model: BenchmarkModelIdentity; started_at: string; ended_at: string; deterministic_checks: BenchmarkDeterministicCheck[]; evidence: BenchmarkEvidenceSummary; completion_decision: string; failure_injections: BenchmarkFailureInjection[]; control_plane: BenchmarkControlPlaneMetrics; economics: BenchmarkEconomics; artifacts: BenchmarkArtifacts; result: BenchmarkResultClassification; } export declare function isComparativeBenchmarkReceipt(v: unknown): v is ComparativeBenchmarkReceipt;