Skip to main content
There is a craft to getting great results out of a coding agent: feed the model the right context at each step, keep its working set small, plan before you build, and prove the work with tests. It has a name, context engineering, and it is most of the difference between an agent that one-shots a feature and one that flails. You do not have to learn it. Smithers bakes context engineering into the agent that drives it. You say what “done” looks like, and the agent assembles the context, the plan, the gates, and the loops on your behalf.

​
Ask your agent for the rundown

The fastest way to understand what is happening under the hood is to ask the agent that runs Smithers for you. Try:
“Give me a rundown of how you do context engineering with Smithers, and what you do when you write a Smithers script.”
It will walk you through its own playbook in plain language and tie it back to your project.

​
What the agent does when it writes a script

When you hand off a real piece of work, the agent is doing this for you:
  • Plans, then validates. It writes a plan with teeth (named tests and a clear definition of done), and puts the gates in before the code, so the result is checked instead of hoped for.
  • Stays in the smart zone. It keeps each step’s context small and focused, with research and planning done up front, so the model spends its attention on the work.
  • Tests end to end. It does not call a feature finished until a real test proves it, and it builds one slice all the way through before starting the next.
  • Delegates the ends and the middle. Strong models plan and review at the two ends; cheaper models handle the routine work in the middle. You get the quality where it matters without paying top rates for everything.