# Claude Code **Claude Code** is Anthropic's agentic coding assistant. It runs directly in your terminal and operates as a full [autonomous agent](autonomous-agent) — reading files, editing code, running shell commands, executing tests, and iterating until a task is complete. ## What Claude Code Can Do - **Read and understand codebases** — Explore files, understand architecture, trace dependencies - **Write and edit code** — Implement features, fix bugs, refactor, write tests - **Run commands** — Execute shell commands, run test suites, build projects - **Use tools** — Call external APIs, search the web, interact with databases - **Connect via MCP** — Use [MCP servers](mcp-server) like AgentRQ to notify humans and await approval ## Claude Code as an Agent Claude Code follows a perceive-plan-act loop. When you give it a task like "add rate limiting to the API," it will: 1. Read the relevant source files 2. Plan the implementation 3. Write the code 4. Run tests to verify 5. Report back — or ask for guidance when needed This makes it fundamentally different from a code completion tool. Claude Code acts on your codebase autonomously. ## Claude Code + AgentRQ By connecting Claude Code to [AgentRQ](/) via [MCP](mcp), you gain real-time visibility into what your agent is doing. Claude Code agents can: - Create [tasks](task) on your [task board](task-board) for work they're undertaking - Send [notifications](notification) when they hit a blocking question - Wait for your [approval](approval) before taking risky or irreversible actions - Continue autonomously once you've replied This is [human-in-the-loop](human-in-the-loop) agent collaboration at the task level. ## Getting Started ```bash # Install Claude Code npm install -g @anthropic-ai/claude-code # Add AgentRQ MCP server to your config # See: https://agentrq.com/docs/getting-started ``` ## Related Terms - [Agent](agent) - [Autonomous Agent](autonomous-agent) - [MCP](mcp) - [Human-in-the-Loop](human-in-the-loop) - [Tool Use](tool-use)