UNPKG

llmverify

Version:

AI Output Verification Toolkit — Local-first LLM safety, hallucination detection, PII redaction, prompt injection defense, and runtime monitoring. Zero telemetry. OWASP LLM Top 10 aligned.

14 lines (13 loc) 737 B
/** * Engine Exports * * @module engines * @author Haiec * @license MIT */ export { HallucinationEngine } from './hallucination'; export { ConsistencyEngine } from './consistency'; export { JSONValidatorEngine } from './json-validator'; export { RiskScoringEngine } from './risk-scoring'; export { LatencyEngine, TokenRateEngine, FingerprintEngine, StructureEngine, BaselineEngine, HealthScoreEngine, isHealthy, getAlertLevel } from './runtime'; export { ClassificationEngine, classify, detectIntent, detectAndRepairJson, evaluateInstructionRules, calculateHallucinationSignals, calculateHallucinationRisk, getHallucinationLabel, calculateCompressionMetrics, calculateCompressionScore, getReasoningLabel } from './classification';