tipcalc/ __init__.py pricing.py # tip(amount, percent) split.py # split_bill(amount, percent, people) tests/ __init__.py test_split.py
Find your way around a codebase
Walking into a big library, you do not read every book. You ask where the maps are, which shelf holds what, and who knows the building. Claude Code can be your guide. You ask it for the map first, and only then pick one shelf to work on.
- Claude Code works in any folder and reads files fresh on each tool call. Start broad ("give me an overview of this codebase"), then narrow ("how is authentication handled?").
- Official prompt recipes to try: "explain the main architecture patterns used here", "what are the key data models?", "trace the login process from front-end to database".
- Type
@and a path to include a file or folder in your message, for example@src/utils/auth.js. A folder reference shows a file listing, not contents. Files over 256KB are not included. - Exploring fills your context with file reads. The docs suggest delegating it: "use a subagent to investigate how our auth system handles token refresh". The subagent reads in its own context window and sends back a summary.
- A specific prompt saves a round of corrections. Name the file, describe the symptom and say what "fixed" looks like. Weak: "fix the login bug". Strong: "users report login fails after the session times out. look at token refresh in src/auth/, write a failing test that reproduces it, then fix it".
- Run
/context allin a session to check which files such as CLAUDE.md were loaded.
- Open a terminal in a project with at least five files (use your own, or any small open-source project). Start
claude. - Type: give me an overview of this codebase. Read the answer.
- Ask: what are the key modules, and how do they depend on each other?
- Pick one file and ask: Explain the logic in @path/to/file (type the real path after the @).
- Ask: where is the code that handles the most important feature, and which tests cover it?
- Write the map in your own words in a notes file: modules, what each does, where the tests are.
- Check: you can explain, without looking, which file you would open to change one named feature.
Videos by other creators on YouTube, for reference. They are not made by or affiliated with The Future Market.
Read the video script
[Host] "Big codebase? Do not read it all. Ask for a map."
[Screen: Claude Code answering 'give me an overview of this codebase'] "Start broad, then narrow down."
[Text on screen: @path/to/file] "Use the at sign to point at one file."
[Screen: a subagent summary] "Let a subagent read many files and bring back only the summary."
[Host] "Your own map in your own words is the real goal."
Pick a project with many files. Use the steps above to write a one-page map: modules, data flow and where the tests live. Then ask Claude to point out one thing in your map that is wrong or missing.
ClaudeStuck? Copy a ready-made prompt for Claude
Copy this, open Claude, paste it, and Claude will walk you through the lesson step by step.
For a grown-up: Claude is for adults (18+). Copy this, paste it into Claude, and use the answer to explain the lesson to your child.
I'm learning "Claude Code Projects: larger codebases, tests and git" on The Future Market. Lesson 1: "Find your way around a codebase".
What the lesson teaches:
- Claude Code works in any folder and reads files fresh on each tool call. Start broad ("give me an overview of this codebase"), then narrow ("how is authentication handled?").
- Official prompt recipes to try: "explain the main architecture patterns used here", "what are the key data models?", "trace the login process from front-end to database".
- Type @ and a path to include a file or folder in your message, for example @src/utils/auth.js. A folder reference shows a file listing, not contents. Files over 256KB are not included.
- Exploring fills your context with file reads. The docs suggest delegating it: "use a subagent to investigate how our auth system handles token refresh". The subagent reads in its own context window and sends back a summary.
- A specific prompt saves a round of corrections. Name the file, describe the symptom and say what "fixed" looks like. Weak: "fix the login bug". Strong: "users report login fails after the session times out. look at token refresh in src/auth/, write a failing test that reproduces it, then fix it".
- Run /context all in a session to check which files such as CLAUDE.md were loaded.
The exercise I'm working on:
Pick a project with many files. Use the steps above to write a one-page map: modules, data flow and where the tests live. Then ask Claude to point out one thing in your map that is wrong or missing.
I'm stuck. Please:
1. Explain the main idea in plain, simple words with one everyday example.
2. Walk me through the exercise one small step at a time. Ask me what I've done so far before giving the next step.
3. Don't just give me the final answer. Help me do it myself, and check my work when I share it.I'm a parent helping my child (about 8 years old) with a lesson called "Find your way around a codebase" from the course "Claude Code Projects: larger codebases, tests and git". Here's what the lesson says: Walking into a big library, you do not read every book. You ask where the maps are, which shelf holds what, and who knows the building. Claude Code can be your guide. You ask it for the map first, and only then pick one shelf to work on. The activity: Pick a project with many files. Use the steps above to write a one-page map: modules, data flow and where the tests live. Then ask Claude to point out one thing in your map that is wrong or missing. Please explain this so an 8-year-old understands: short sentences, a fun everyday example, and no hard words. Then give me 3 simple questions I can ask my child to check they understood, and a way to do the activity together, one tiny step at a time.
Sources
- Claude Code docs, Common workflows (Understand new codebases, Reference files and directories, Delegate research to subagents): code.claude.com/docs/en/common-workflows
- Claude Code docs, Best practices (Write a specific prompt): code.claude.com/docs/en/best-practices
From idea to spec to plan
This lesson is part of full access: every course and every level for one payment.
Tests are the loop
This lesson is part of full access: every course and every level for one payment.
Git as your safety net
This lesson is part of full access: every course and every level for one payment.
Review code you did not write
This lesson is part of full access: every course and every level for one payment.
Subagents and parallel work
This lesson is part of full access: every course and every level for one payment.
Stay in control on bigger jobs
This lesson is part of full access: every course and every level for one payment.
Final project
This lesson is part of full access: every course and every level for one payment.
Quiz
How do you include one file in your message to Claude Code?
Type @ and the path, for example @src/utils/auth.js.
Why ask Claude to interview you and write SPEC.md?
It turns a loose idea into a written definition of done before any code is built.
After two failed corrections, what do the docs suggest?
Run /clear and write a better prompt instead of a third correction.
What makes a good check for Claude to run?
It comes back pass or fail, such as tests, a build or a linter.
What does a Stop hook that exits with code 2 do?
It sends Claude back to work, with the printed reason, instead of letting it finish.
Do checkpoints replace Git?
No. They cover Claude's own file edits in a session; Git is your real history.
What does
claude --worktree feature-authdo?Starts a session in a new git worktree on its own branch, so parallel work does not collide.
Why use a reviewer that did not write the code?
It is not swayed by the reasoning that produced the code and sees problems the writer misses.
How can a reviewer subagent be made unable to edit files?
List only read tools, such as Read, Grep and Glob, in its tools field.
What is the main cost of subagents?
Tokens, because each subagent sends its own requests and counts against the same usage.
Why put a never-rule in a deny rule or hook instead of only CLAUDE.md?
CLAUDE.md is guidance Claude usually follows; deny rules and hooks are applied by Claude Code itself.
Which commands help control a long session's cost and context?
/clear between tasks, /compact with a focus, and /usage to see where usage goes.
Final exam
This lesson is part of full access: every course and every level for one payment.
Certificate
This lesson is part of full access: every course and every level for one payment.