Loop Engineering with Claude
Learn the loop, not the tool. A practical workshop on setting up context, choosing tools, and building a repeatable cycle of making and checking.
The question
How do you turn an AI coding session into a workflow you can repeat and trust?
The approach
Connect project memory, tools, skills, and verification in a loop. Give the work a clear goal and a condition for stopping.
Why it matters
The setup and habits carry forward as tools change. The useful unit is the workflow, with evidence that the result works.
The workshop deck
Use the buttons, or focus the deck and use the arrow keys. On a phone, a landscape view makes the slides easier to read.
What the workshop covers
Set up the context
Project memory, permissions, tools, skills, subagents, and hooks give a session structure. The deck walks through the setup and when a command-line tool is a better fit than an MCP connection.
Keep knowledge alongside the work
Separate reusable knowledge from implementation details. Preserve what the next session needs, and turn repeated tasks into skills.
Build, check, and repeat
Choose a loop that fits the task. Define what success looks like, check the result, and stop when the condition is met.