47% OFFYearly Pro
$30/mo$16/mobilled yearlyGet Pro
Free ebookCoding AgentsClaude CodeHarness Engineering

Production coding agents: a harness engineering guide

A harness engineering guide for production coding agents. The streaming loop, permission-governed tool dispatch, context management, sandboxing, and eval. Grounded in the claude-code-from-scratch reverse-engineering.

Inside the ebook

What you get

  • Wire a streaming agent loop where the model is the only source of decisions
  • Design permission-governed tool dispatch that is safe by default
  • Manage context across long sessions with compaction and memory tiers
  • Sandbox shell execution with timeouts, resource limits, and typed IO
  • Evaluate the agent on tasks that reflect real user intent

Inside

  • What harness engineering means (and why it matters)
  • The streaming agent loop
  • Permission-governed tool dispatch
  • Context management and compaction
  • Sandboxed shell execution
  • Memory tiers: session, project, skill
  • Evaluating a coding agent
  • Ship checklist
Checking access…

Prefer a walkthrough? Watch the companion webinar.

Keep going

What to do next

Go deeper

Learn to Build your own Coding Agent (Claude Code)

Reverse-engineer how Claude Code works. Then build your own production AI coding agent from scratch.

Open course