Claude Code Practical Scenarios and Workflows

These are the most common and practical Claude Code workflows, covering everyday tasks like taking over a new project, fixing bugs, refactoring, adding tests, and writing documentation.

Common Workflows

1. Understanding a new codebase: after entering the project directory, run claude and start by asking "Give me an overview of the codebase", "Explain the main architectural patterns", or "What are the key data models".
2. Fixing bugs efficiently: paste the full error and reproduction steps directly, e.g. "I'm seeing this error when running npm test — please help me find and fix the problem".
3. Refactoring: first have Claude locate legacy APIs or duplicated logic, then explicitly ask it to refactor while "keeping the existing behavior unchanged".
4. Handling tests: have Claude identify uncovered code first, then add tests and edge cases, and finally run the tests to verify.
5. Creating a Pull Request: summarize the changes first, then ask for a PR title, summary, and test notes, adding security notes when necessary.
6. Handling documentation: ask Claude to find functions missing comments and add JSDoc, README snippets, or migration notes.
7. Using images: drag and drop screenshots, design mockups, or error images — great for debugging UI issues and reproducing visual designs.
8. Setting up project memory: use /init to generate CLAUDE.md, recording common commands, architectural conventions, and coding standards.
9. Unix pipe integration: use claude -p to consume logs, build output, and diffs — well suited for scripting and CI pipelines.
10. MCP server configuration: use claude mcp add/list/get/remove to manage external capability extensions.

MCP Server Configuration Example

claude mcp add my-server -e API_KEY=123 -- /path/to/server
claude mcp list
claude mcp get my-server
claude mcp remove my-server

Pipe Integration Example

cat build-error.txt | claude -p "concisely explain the root cause of this build error" > output.txt