/** * AAL (Agent Action Lock) - Semantic Core * * @remarks * This is the semantic core of AAL. It only contains: * - Pure functions (stateless) * - Immutable value objects * - Types * - Decision processing * * **Architecture:** * - Consumes ISLSignal (not ISLResult) to maintain layer separation * - Applies configurable policies (ALLOW/WARN/BLOCK) * - Builds remediation plans (what to do, not how; SDK/AI agent performs cleanup) * - Does not execute actions (that is SDK responsibility) * * **Does NOT contain:** * - Prompt injection detection (goes to ISL) * - Stateful services (go to SDK) * - Action execution (goes to SDK) */ export { createAnomalyScore, isHighRisk, isLowRisk, isWarnRisk, isRoleProtected, isContextLeakPreventionEnabled, isInstructionImmutable, isIntentBlocked, isScopeSensitive } from './value-objects/index.js'; export type { AnomalyScore, PolicyRule } from './value-objects/index.js'; export { resolveAgentAction, resolveAgentActionWithScore, buildDecisionReason, buildRemediationPlan, validateAgentPolicyThresholds } from './process/index.js'; export type { DecisionReason } from './process/index.js'; export { buildAALLineage } from './lineage/index.js'; export { ACTION_DISPLAY_COLORS, getActionDisplayColor } from './constants.js'; export type { AnomalyAction, RemediationPlan, BlockedIntent, SensitiveScope, ProtectedRole, ImmutableInstruction, AgentPolicy, } from './types.js'; //# sourceMappingURL=index.d.ts.map