DeepLearning.ai-Courses-ClaudeCode system prompt

--- title: Claude Code A Highly Agentic Coding Assistant DeepLearning.AI slug: claude-code-a-highly-agentic-coding-assistant-deep-1755233480230 source: https://learn.deeplearning.ai/courses/claude-code-a-highly-agentic-coding-assistant/lesson/hhfj3/prompts-&-summaries-of-lessons datetime: 2025-08-15T04:51:20.229Z --- # Prompts & Summaries of Lessons This note includes the links to the prompts used in the lessons, additional resources and a summary of Claude Code features. **Note**: To mark this reading item as complete, make sure to scroll down to the end and click on "Mark as Complete". ## Prompts Here are the links to the lessons' notes and prompts: - [Prompts of Lesson 2: Setup & Codebase Understanding](https://github.com/https-deeplearning-ai/sc-claude-code-files/blob/main/reading_notes/L2_notes.md) - [Prompts of Lesson 3: Adding Features](https://github.com/https-deeplearning-ai/sc-claude-code-files/blob/main/reading_notes/L3_notes.md) - [Prompts of Lesson 4: Testing, Error Debugging and Code Refactoring](https://github.com/https-deeplearning-ai/sc-claude-code-files/blob/main/reading_notes/L4_notes.md) - [Prompts of Lesson 5: Adding Multiple Features Simultaneously - Using Git Worktrees](https://github.com/https-deeplearning-ai/sc-claude-code-files/blob/main/reading_notes/L5_notes.md) - [Notes for Lesson 6: References to GitHub Integration & Hooks](https://github.com/https-deeplearning-ai/sc-claude-code-files/blob/main/reading_notes/L6_notes.md) - [Prompts of Lesson 7: Refactoring a Jupyter Notebook & Creating a Dashboard](https://github.com/https-deeplearning-ai/sc-claude-code-files/blob/main/reading_notes/L7_notes.md) - [Prompts of Lesson 8: Creating Web App based on a Figma Mockup](https://github.com/https-deeplearning-ai/sc-claude-code-files/blob/main/reading_notes/L8_notes.md) ## Additional Resources To learn more about these features as well as other features, you can check: - [Claude Code Documentation](https://docs.anthropic.com/en/docs/claude-code/overview) - [Claude Code Common Workflows](https://docs.anthropic.com/en/docs/claude-code/common-workflows) - [Claude Code Best Practices](https://www.anthropic.com/engineering/claude-code-best-practices) - [Claude Code Use Cases](https://www.anthropic.com/news/how-anthropic-teams-use-claude-code) There's also a great course on Anthropic Academy that you can check out to see more examples with Claude Code: - [Claude Code in Action](https://anthropic.skilljar.com/claude-code-in-action) ## Summary of Claude Code Features ### Managing Project Memory - `/init`: Claude Code scans your codebase and creates CLAUDE.md file inside your project directory. - CLAUDE.md guides Claude through your codebase, pointing out important commands, architecture and coding style. It's automatically included in the context each time you launch Claude Code. - Here's an [example](https://github.com/https-deeplearning-ai/ragchatbot-codebase/blob/main/CLAUDE.md) of a CLAUDE.md file generated by `init` for the RAG chatbot example. - `#`: Use `#` to quickly add a memory. Useful when you see Claude Code repeats an error. - **Example 1**: since the project is a `uv` project, we added these to CLAUDE.md file using `#`: - `#`use uv to run python files or add any dependencies - **Example 2**: you can inform Claude Code about the database schema, in this case since you have a vector database, you can inform Claude Code about the collections stored in the vector database: - `#`The vector database has two collections: - `course_catalog`: - stores course titles for name resolution - metadata for each course: title, instructor, course\_link, lesson\_count, lessons\_json (list of lessons: lesson\_number, lesson\_title, lesson\_link) - `course_content`: - stores text chunks for semantic search - metadata for each chunk: course\_title, lesson\_number, chunk\_index ### Summary of Claude Code Commands | Command | Description | | --- | --- | | `/clear` | clears current conversation history | | `/compact` | summarizes current conversation history | | `ESC` | interrupt Claude to redirect or correct it | | `ESC ESC` | rewind the conversation to an earlier point in time | | `@` | Mention files with `@` to include its content in your request | | `/mcp` | Manage MCP connection & check available MCP servers with their provided tools ([MCP with Claude Code](https://docs.anthropic.com/en/docs/claude-code/mcp)) | You can use regular bash command within Claude Code, but you need to start the command with `!` (for example: `!pwd`). You can type `exit` to quit Claude Code. | Shortcuts | Description | | --- | --- | | `shift`+`tab` | switch between planning and auto-accept mode | | take a screenshot | `cmd`\+ `shift`\+ `ctrl` + `4` (Mac) or `Win` + `Shift` + `S` (Windows) | | paste a screenshot | `Ctrl` + `V` (might not work on Windows) | ### Additional Claude Features - **Extended Thinking Mode** For complex tasks (e.g., complex architectural changes, debugging complicated issues), you can use the word "think" to trigger [extended thinking mode](https://docs.anthropic.com/en/docs/claude-code/common-workflows#use-extended-thinking). There are several levels of thinking: "think" < "think hard" < "think harder" < "ultrathink." Each level allocates more thinking budget for Claude. - **Use of subagents** You've learned that one of the out-of-the-box tools for Claude Code is **Task**, which Claude Code can use to launch subagents for complex multi-step tasks. You can explicitly ask Claude Code to use subagents to brainstorm ideas or to investigate multiple aspects of a question or a problem you want to solve. These built-in agents are of general purpose. You can also create your customized specialized subagents. Each subagent has its own context window, and you can define a custom system prompt and specific tools for each subagent. This part is not covered in this course, but you can check the details in the documentation [here](https://docs.anthropic.com/en/docs/claude-code/sub-agents).