Package com.shaft.pilot.ai


package com.shaft.pilot.ai
  • Class
    Description
    Per-request limits for model usage.
    Provider capabilities that callers can inspect before submitting evidence.
    Enforces approval, redaction, budgets, concurrency, retries, schema validation, and circuit breaking before accepting provider output.
    Immutable image evidence.
    SHAFT-owned contract implemented by optional AI providers.
    Safe provider availability result.
    Resolves explicit current-thread providers before ServiceLoader providers, then falls back to the disabled provider.
    Immutable provider-neutral request.
    Fluent builder for AiRequest.
    Provider-neutral AI result with deterministic fallback.
    Normalized provider and policy outcomes.
    Provider-reported usage metadata.
    Explicit consent for local or remote inference and evidence categories.
    Deterministic no-op provider used when AI is disabled or unavailable.
    Evidence categories used by approval and redaction policies.
    Textual evidence supplied to a provider after approval and redaction.
    Describes where a provider processes submitted evidence.