mlld:agents

Design and deploy modular LLM agents with @meta and @taskPrompt exports.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/mlld-lang/mlld --skill mlld-agents
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mlld:agents
Source: https://github.com/mlld-lang/mlld/tree/main/plugins/mlld/skills/agents
Command: npx skills add https://github.com/mlld-lang/mlld --skill mlld-agents

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

mlld agents provide a structured way to build diverse AI agents (tool, event-driven, and workflow) that can be deployed and governed in a consistent framework.

Core Features & Use Cases

  • Convention-based agent modules that export @meta and prompt templates like @taskPrompt for routing and dispatch.
  • Supports three paradigms: Tool Agents (MCP tools), Event-Driven Agents (routers/dispatchers), and Workflow Agents (stateless jobs), with guards and model-selection guidance.
  • Includes examples and best practices for securing prompts, scoping tools, and integrating with the llm/ directory structure.

Quick Start

Create a new agent module under llm/agents following the exported @meta and @taskPrompt pattern, then run mlld to load and test it.

Frequently Asked Questions about mlld:agents

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build secure LLM agents for automated workflows?

Build secure LLM agents by using convention-based modules that export @meta and prompt templates like @taskPrompt. This framework enforces security guards and model guidance to safely deploy tool, event-driven, and workflow agents.

What is the best way to structure event-driven routers and dispatchers for AI agents?

Structure event-driven routers and dispatchers by creating agent modules under the llm/agents directory. Export @meta and @taskPrompt patterns to handle routing logic, apply security guards, and guide model selection within your automated workflows.

How do I create MCP tool agents with scoped security guards?

Create MCP tool agents by following the convention-based contract that exports @meta and @taskPrompt. Apply built-in security guards and scoping practices to restrict tool access and secure prompts during automated AI workflow execution.

Can I run stateless workflow agents as background jobs using this convention?

Yes, you can run stateless workflow agents as background jobs. The framework supports three paradigms—Tool Agents, Event-Driven Agents, and Workflow Agents—enabling stateless job execution with appropriate model guidance and security guards.

Does mlld support deploying modular LLM agents with different execution paradigms?

mlld supports deploying modular LLM agents across three paradigms: Tool Agents for MCP tools, Event-Driven Agents for routing, and Workflow Agents for stateless jobs. It uses a consistent convention-based contract for all types.

Why do I need convention-based contracts like @meta and @taskPrompt for AI agents?

Convention-based contracts like @meta and @taskPrompt provide a structured way to govern diverse AI agents. They ensure consistent deployment, enforce security guards, and standardize prompt routing across tool, event-driven, and workflow agents.