Search

Evaluation harness — scenario × model matrix, deterministic / LLM judges, `pass^k` reliability summaries, and JSON/Markdown reports. Includes the `agent-engine-eval` CLI. Runtime-optional; for release gates and model selection.

PostgreSQL-backed long-term `MemoryManager` — scoped retrieval, provenance, and supersede/tombstone lifecycle — through a host-injected pg-compatible client.

Cut your LLM prompt size by 40-70% in one line of code -- semantic chunking + extractive summarization that preserves meaning, instructions, and key entities.

Run any agent (Claude, Codex, custom) on any machine — with no API key on the machine. A secure, self-hosted proxy for models and tools.

Install AI agent skills for Claude Code, Codex, Antigravity, and OpenCode

Thin HTTP client for llmuxd — send prompts to running AI agent CLI sessions over REST

Free, open-source, async-first web scraping toolkit for LLM agents

Production LLM calls. Just the three lines. Reliability, native caching, and reversible context compression on by default.

Drop-in LangChain JS tools for Agent402: turn 1,000+ pay-per-call web tools into DynamicStructuredTool instances, with auto-payment (proof-of-work free, USDC via x402 paid). Works with LangGraph, agents, and LCEL chains. Zero deps.

Autourgos LLM wrapper for the OpenAI Chat Completions API

Wrap AI agent tool functions with dry-run preview and user confirmation before execution

Universal TUI agent for autonomous research and writing, powered by llama.cpp

Forecast which agent calls burn Anthropic Claude Agent SDK credit after the 2026-06-15 billing change — Claude Code / claude-code-action / claude -p cost, monthly risk, and cheaper alternatives.

Multi-agent software engineering plugin for OpenCode — implements Stages I-IV from "The End of Software Engineering" (arXiv:2606.05608)

AgenticUptime CLI - infrastructure monitoring for coding agents. Add servers, check health, view events, conductor AI analysis. As easy as git.

AI-assisted live staging QA layer for Playwright test suites. Parses @qa-scenario annotations from *.spec.ts into structured QA scenarios, then uses Hermes Agent to log into staging, visit target pages, and judge live DOM against test intent — returning p

Read-only LangChain tools for RustChain — let any agent query the hardware-attested agent economy.

Self-contained ReAct (Reasoning + Acting) agent for the Autourgos framework — works with any OpenAI-compatible LLM

Credential primitives for the P3 Agentic Registry — types, canonicalization, hashing, and validation

<div align="center"> <img src="public/logo.svg" alt="Skill Atlas Logo" width="120" /> <h1>Skill Atlas Workbench</h1> </div>