Spec-driven coding agent framework: knits tool dispatch and model calls into an agent loop for any OpenAI-compatible endpoint.
An AI agent framework that builds itself — multi-model ReAct + tiered context engine + multi-agent + XML workflows, zero LangChain
Open-source antivirus and runtime protection for AI agents, tools, MCP servers, plugins, skills, and package installs.
Score your AI agent's replies for human-likeness before sending — free trial, then pay-per-score via x402 (USDC on Base, no account, no API key). Includes LangChain and CrewAI tools.
Open-source antivirus and runtime protection for AI agents, tools, MCP servers, plugins, skills, and package installs.
