2026 version: OpenClaw MCP
Extreme usecase of Model Context Protocol: allowing an agent access to your full computer, keyboard and password manager to perform tasks in your name. Communication is done through chat, commands are given but the agent performs the task and decides on what steps to ask for human intervention.
Gigantic security implications: agents have full control over everything including mailboxes, passwords and creditcards and are inherently sensitive to Prompt Injection.
Advanced technically adept users understand that you need to buy separate hardware, telephones and accounts to attempt to sandbox this kind of technology, but the majority will just install the software in search of productivity gains and woefully accept the security risks, and will only regret using it when damages have already happened.
2025 version: n8n workflows
Exciting new technology in which classic no-code like workflows are created but glued together with use of (local) LLM’s. These can be triggered by events or by recurring schedules, setting the first steps towards truly autonomous agents capable of carrying out responsibilities and tasks
Typical Agent programming consists of a combination of the LLM engine, a query, tools and memory
agent tooling
- LangGraph from LangChain;
- Amazon Bedrock’s AI Agent framework;
- Rivet, a drag and drop GUI LLM workflow builder; and
- Vellum, another GUI tool for building and testing complex workflows
tooling
- N8n - ifttt / zapier / node-red like no-code environment with strong LLM integration, also including local LLMS like ollama
sources
Link to original