Agent Engineering

The practical craft of building agents that survive contact with real work: tool design, context, evaluation, failure modes and the boundaries worth enforcing.

A secure machine hall where independent AI capability modules connect through guarded interfaces while one module is safely replaced
Artificial IntelligenceArchitectureAgent Engineeringcomposable-systems
Composable AI Systems: Architecture That Can Change Without Losing Control
A practical architecture for building AI systems from bounded capabilities, explicit contracts, least-privilege tools, durable state, and verification gates.
Published on
A volatile luminous intelligence being fitted into a precise mechanical harness inside an experimental engineering bay
Artificial IntelligenceAgent EngineeringArchitectureOrchestration
The Agent Harness Is the Real Product
Why model capability only becomes dependable work through a harness that manages context, tools, state, permissions, recovery, and verification.
Published on
An autonomous machine moves around a circular workshop track, completing inspection and verification stations before the path opens toward a bright finish gate
Artificial IntelligenceAgent EngineeringcodexOrchestration
The Loop and Goal Workflow: How Long-Running AI Work Actually Finishes
A practical look at Codex Goal mode, automatic continuation, verification, steering, and the discipline required to make long-running agent work finish honestly.
Published on
A dark research observatory containing three celestial model cores, a protected black-glass intelligence core, and a vast open lattice of expert modules
Artificial IntelligencellmAgent Engineering
Sol, Fable, Mythos, and Kimi K3: The AI Model Race Is Splitting Apart
What the newest frontier models reveal about tiered intelligence, long-running agents, safety-gated capability, and the return of open-weight competition.
Published on
Warm engineering desk with a terminal workstation, notebooks, sketches, and dawn light across a focused coding setup
Artificial IntelligenceAgent EngineeringClaude Code
A Practical Start With Claude Code
A working introduction to Claude Code that focuses on memory, operating rhythm, and the habits that make it useful on real projects.
Published on
Lamplit tabletop atlas with glowing routes, miniature terminals, relay towers, and symbolic landmarks for an AI engineering journey
Artificial IntelligenceAgent EngineeringClaude Codemcp
Agent Engineering Playbook
A guided series on Claude Code, skills, MCP, subagents, orchestration, recursive workflows, and Gas Town.
Published on
Colorful checkpoint scene where a central blue figure mediates guarded handoffs between outside systems, couriers, and controlled gateways
Artificial IntelligenceAgent EngineeringmcpClaude Code
MCP From First Principles
A practical explanation of the Model Context Protocol, why it matters, and how to connect Claude Code to external systems without treating it like a black box.
Published on
Mission-control room overlooking bright work lanes that flow through checkpoints toward a final integration gate
Artificial IntelligenceAgent EngineeringOrchestrationmulti-agent
Orchestration Patterns for Real Agent Work
A practical guide to planner-executor-reviewer loops, handoffs, verification gates, and the operating patterns that make multi-agent work reliable.
Published on
Surreal sequence of nested frames and rooms receding inward as a central figure looks through recursive layers
Artificial IntelligenceAgent EngineeringrlmOrchestration
RLMs, Context Rot, and Recursive Orchestrators
What Recursive Language Models are, why long-context systems degrade, and how recursive decomposition changes the design of agent orchestration.
Published on
123
Page 2 of 3
HomeAboutArticlesProjectsChat