Universal library for evaluating AI models

The LLM Evaluation Framework for TypeScript

OpenCode Goal Mode 0.9 — goal agent + server plugin: goal_set, idle evaluator, auto-continue until YES. Simple, no sidebar, no guard.

A mirror of Facebook's Warning

Simulate evaluation of AWS IAM policies

Check conditional statements and return true/false

Test/optimize loop for agent skills — run spec'd scenarios on pi, LLM-judge, score, review, re-run

AWS SDK for JavaScript Ec2 Client for Node.js, Browser and React Native

Rulvar evals: eval cases, golden outputs, rubric and judge graders, matrix sweeps, canary fingerprint.

TypeScript definitions for warning

<p align="center"> <h2 align="center">Runner plugin for Vite</h2>

Agent Looper — repo-agnostic fix-until-green harness (Cursor / Cline / OpenCode / Pi / Codex; configurable reviewRuntime judge)

Invariant and non-null/undefined assertions

Efficient Binary heap (priority queue, binary tree) data structure for JavaScript / TypeScript. Includes JavaScript methods, Python's heapq module methods, and Java's PriorityQueue methods.

Evaluate LLM and rule-based judges with mutation cases whose generation does not depend on the judge under test.

[email protected]:stalniy/ucast.git

AI/LLM judge matchers for Playwright — toPassRubric/toScoreAtLeast/toMatchImage over Ollama, OpenAI-compatible endpoints (OpenAI/OpenRouter/NVIDIA/Groq/…), and native Claude

Helper functions for assembling CloudFormation templates in JavaScript

Retrieval latency ladder benchmarks + CI regression gates for @remnic/core

Return a function which tests if every element in an array-like object passes a test condition.