Constellation · First Principles

Foundations: Craft that outlives any one tool

Seven tool-agnostic guides to the craft beneath every coding agent: the named agentic design patterns, context and prompt engineering, building an MCP server and the protocol behind it, agent security and sandboxing, spec-driven development, and evals that prove your setup actually works.

7 of 7 guides live · fact-checked against official docs

Chapter 0

Agentic Design Patterns

Learn the tool-agnostic patterns — the agent loop, chaining, routing, fan-out, orchestrator–workers, reflection — and exactly when each one wins.

12 minagentic-patterns / architecture / first-principles / agent-loop / orchestration

Chapter 1

Context Engineering

Treat the context window as a finite budget and learn the tool-agnostic craft that makes every coding agent sharper.

11 mincontext-engineering / prompt-engineering / foundations / mcp / agents

Chapter 2

Build an MCP Server

Write a working MCP server in TypeScript or Python, test it with the Inspector, and register it with your agent.

11 minmcp / sdk / tools / server / python / typescript

Chapter 3

MCP Protocol

Read the Model Context Protocol itself — architecture, primitives, transports, and auth — so every tool's MCP screen becomes a thin skin over machinery you understand.

11 minmcp / protocol / foundations / architecture / oauth / transports / sdk

Chapter 4

Security & Sandboxing

Break the lethal trifecta and wrap deterministic, OS-level boundaries around any coding agent.

12 minsecurity / prompt-injection / sandboxing / mcp / permissions / ci

Chapter 5

Spec-Driven Dev

Write the spec first, let the agent plan, slice, and implement from it — and stop fixing confidently-wrong code.

11 minspec-driven-development / spec-kit / kiro / plan-mode / workflow / foundations

Chapter 6

Evals for Agents

Build a repeatable eval loop so every prompt, tool, and model change becomes a measurement instead of a leap of faith.

11 minevals / testing / llm-as-judge / reliability / agents