agent-native-architecture

Codify prompt-native agent architecture with prompt-defined outcomes and primitive tools.

1|1|Updated Jan 2, 2026

Search Tags:#architecture #context #workspace #tools #parity #prompt-native #agent-native

Ready to run

npx skills add https://github.com/nbbaier/compound-engineering-amp --skill agent-native-architecture-nbbaier

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-native-architecture
Source: https://github.com/nbbaier/compound-engineering-amp/tree/main/skills/agent-native-architecture
Command: npx skills add https://github.com/nbbaier/compound-engineering-amp --skill agent-native-architecture-nbbaier

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill codifies the Prompt-Native philosophy: outcomes are defined in prompts, features are prompts, and primitives are tools. It guides the development lifecycle, self-modification, and when not to push behavior into code.

Core Features & Use Cases

The Prompt-Native Philosophy: define outcomes in prompts; agents figure out HOW to achieve them
Features are prompts, not code39
Tools are primitives that enable capability, not workflows
Development lifecycle: start in prompt, iterate rapidly, graduate to code when stable
Self-modification patterns with explicit guardrails

Quick Start

Build a prompt-native agent by defining primitive tools, articulating behavior in the system prompt, and letting the agent determine HOW to achieve outcomes.