Claude Code Mastery: Subagents, MCP, Hooks & AI Workflows

Examine slash commands in cloud code that use markdown prompts to automate workflows, save time, and standardize team processes. Learn project-based versus user-scoped commands and built-in versus custom options.

Explore the structure of a slash command, starting with the forward slash and command name, then use positional arguments like 123 and high priority to define tasks.

Learn to pass multiple inputs to slash commands using positional parameters, demonstrated by a review file that prints a PR with a priority and assigns it to an author.

Explore frontmatter metadata for advanced slash commands, including allow tools, argument hints, description, model, and disabled model invocation, with hands-on examples like expense tracker doc.

Learn how bash command execution with an exclamation prefix runs before slash commands, returning output in the command context, guiding PR status checks, branching, and a single git commit.

Use a slash command to generate and maintain docs for a bash expense tracker command-line interface, including readme templates, docstrings, and coverage checks to keep docs in sync with code.

Explore a hands-on slash command for generative AI-assisted development that refactors a Python expense tracker into an object-oriented, test-driven codebase with clear before-and-after comparisons.

Walk through a hands-on git workflow using a slash command for a quick commit, refactoring, testing, updating the readme, and pushing to GitHub with an expense tracker project.

Learn to use slash commands with MCP, enabling dynamic discovery of prompts from a connected MCP server. Create weather queries with MCP_weather_get_forecast using latitude and longitude.

Build an end-to-end expense tracker workflow that automatically refactors code and generates up-to-date documentation using slash commands and sequential execution.

Explore subagents, specialized AI assistants with defined roles and isolated context windows that preserve context, enable parallel work, and support reusable, secure workflows across multiple projects.

Explore how Claude Code enables an ai-assisted travel activity planner agent, with research, personalization, and activity curation for destinations like Tokyo and Costa Rica, guided by examples and quality standards.

Run the travel activity planner agent with Claude Code, perform web searches for New York outdoor activities, and generate a four-day itinerary with weather and top recommendations.

Create a travel activity planner agent on cloud code with slash command agents, set it in user root directory, and configure tools and system prompt to generate activities and itineraries.

Create a restaurant scout agent to find top Tokyo options, using the sonnet model with an orange background, guided by a system prompt and parallel tool use.

Learn how to run multiple Claude agents in parallel, sequentially, or in a hybrid, and compare performance for planning a New York family itinerary.

Run agents sequentially to create a travel itinerary for New York City and then find restaurants in the proposed areas, illustrating sequential execution for a family trip.

Create a project-level subagent to review and design an API for the expense tracker, establishing RESTful and GraphQL best practices, security, and documentation for browser access.

Hands-on with an api reviewer agent to analyze an expense tracker, design a restful api architecture, and implement a fast api server with authentication, middleware, error handling, and testing utilities.

In Claude Code: generative ai-assisted development, this hands-on API review implements a RESTful API for an expense tracker, detailing endpoints, docs, and JWT authentication with rate limiting.

Orchestrate an advanced workflow that combines an API reviewer agent and a slash command to review, refactor, test, and push code to GitHub for the expense tracker.

Understand slash commands as quick, template-based prompts stored in markdown files for instant tasks in Claude, and how subagents enable isolated, parallel processing and context preservation.

Compare slash commands and subagents to choose the right tool for your workflow. Slash commands enable quick, repetitive tasks within one context, while subagents provide isolated, parallel research and verification.

Preserve context across coding sessions to maintain coherence, enabling iterative edits, cross-file awareness, and retrieval augmented memory for a persistent, intelligent coding partner.

Explore cloud code memory, a markdown-based system that persists context across sessions and organizes memory into enterprise, project, user, and conversational layers for consistent, efficient development.

Manage memory on the fly for temporary session context using the pound symbol to apply camel case to variables and functions; Use memory commands to view project and user memory.

Explore how user memory, project memory, and root Claude memory determine preferences and precedence, with explicit overrides guiding non-negotiable project standards and personal workflows.

Explore Claude memory hierarchy: direct memory messages, idea selection, markdown files (subdirectory over root); user settings; project memory overrides user memory as team source of truth in version control.

Explore memory models in cloud code through a Python CLI finance tracker, illustrating project, user, and system memory precedence and its impact on data persistence and no external dependency design.

Describe how a user memory guides building a Python CLI project with click, decimal money handling, and test_ prefixed tests; the setup creates a finance tracker and runs tests.

Investigate project level memory in a Python CLI finance tracker using cloud markdown files and environment. See how memory precedence favors project memory over user memory and float over decimal.

Create a subdirectory memory for a finance tracker api using Claude memory, overriding routes and conventions in a FastAPI app with single quotes, logger-based error handling, and pydantic models.

Explore client-server architecture, including the roles of clients, servers, and protocols like HTTP and REST, and see how MCP standardizes APIs to enable scalable ai workflows.

Explains how large language models evolve from prediction to integrating external tools, via RAG and AI agents, addressing limitations and scaling toward MCP.

Explore MCP, the model context protocol that standardizes two-way communication between MCP clients and servers, enabling client apps to access external services through a translator and leverage capabilities.

Explore the three MCP server components: tools, resources, and prompts—and see how AI performs actions, reads data, and follows user-defined prompts for tasks and workflow automation.

Explore MCP transport types—from stdio for local runs to HTTP with SSE, and streamable HTTP for cloud native, stateless communication—and learn how to choose infrastructure for scalable AI tools.

Explore MCP transport: streamable HTTP enables stateless, chunked requests for cloud-native serverless deployments, supporting enterprise scalability and letting AI models interact with external environments from local to global.

Explore MCP transport types and how they enable client-server communication, from stdio local input and output streams to http with sse for streaming responses and http post for sending messages.

Explore a real-world agent system architecture for an excel document analyzer, featuring Claude Desktop as the client, a server with tools, prompts, and file system access.