agentcarousel 0.7.0

Unit tests for AI agents. Run behavioral tests in CI, score with an LLM judge, and export signed evidence your auditors accept.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

//! Pluggable **evaluators** that score a finished [`crate::CaseResult`] against fixture rubrics
//! or external references: [`RulesEvaluator`], [`GoldenEvaluator`], [`ProcessEvaluator`],
//! [`JudgeEvaluator`], and the [`Evaluator`] trait.

mod assertions;
mod golden;
mod judge;
mod process;
mod rules;
mod trait_def;

pub use golden::{
    evaluate_for_promotion, GoldenEvaluator, PromotionMeta, PromotionResult,
    PROMOTE_CRITICAL_THRESHOLD, PROMOTE_EFFECTIVENESS_THRESHOLD,
};
pub use judge::{run_prompt_audit, JudgeEvaluator};
pub use process::ProcessEvaluator;
pub use rules::{evaluate_case, RuleEvaluation, RulesEvaluator};
pub use trait_def::{Evaluator, EvaluatorError, EvaluatorKind};