What is Claude Code? 2026 Comprehensive Turkish Guide: Setup, Hooks, MCP, Sub-Agents
A zero-to-advanced Turkish guide for Anthropic's terminal-native agentic code assistant Claude Code: installation (npm/Homebrew), CLAUDE.md file, slash commands, MCP server integration (GitHub, Postgres, Linear), hooks (PreToolUse, PostToolUse, Stop), sub-agents, IDE integration (VS Code, JetBrains, Neovim), cost optimization, KVKK compliance. 15 practical commands + 8 use cases.
One-line answer: Claude Code is the terminal-native agentic AI coding assistant — IDE-agnostic, MCP-integrated, sub-agent-capable, and the strongest Turkish-fluent paid option for Turkish developers.
- Claude Code is Anthropics terminal-native agentic AI coding assistant, released Feb 2025. Direct access to Claude Opus 4, Sonnet 4.6, Haiku 4.5.
- No bundled IDE — runs alongside VS Code, JetBrains, Neovim, Cursor — fully IDE-agnostic.
- Three core powers: (1) agent loop for multi-step task execution, (2) MCP for native integration with GitHub/DB/Linear/etc., (3) sub-agents via Task tool for parallel exploration.
- Install in 60 seconds: npm install -g @anthropic-ai/claude-code, then claude login. Node 18+. Mac/Linux/Windows.
- CLAUDE.md is the per-project instructions file. Tech stack, git policy, security rules, file layout go here.
- Hooks (PreToolUse, PostToolUse, Stop) define custom workflows: pre-commit tests, KVKK regex enforcement, etc.
- Cost: Claude Pro ($20/mo) for typical use. Heavy: Claude Max ($100-200/mo) or direct Anthropic API ($3/$15 Sonnet, $15/$75 Opus per 1M tokens).
1. What is Claude Code?
Anthropic's terminal/CLI agentic AI coding assistant launched Feb 2025. Direct access to Claude models with MCP, sub-agents, hooks, and IDE-agnostic architecture.
2. Installation
npm install -g @anthropic-ai/claude-code then claude login. Requires Node 18+, supports macOS/Linux/Windows.
3. CLAUDE.md
A project-root markdown file that Claude Code reads every session. Contains tech stack, git policy, security rules, file layout, conventions. Hierarchy: global (~/.claude/CLAUDE.md), project (./CLAUDE.md), local (./CLAUDE.local.md).
4. Slash Commands
Built-in: /init, /clear, /compact, /mcp, /hooks, /cost, /model, /permissions, /agents. Custom slash commands defined in ~/.claude/commands/*.md.
5. MCP — Model Context Protocol
Anthropic's open protocol for connecting LLMs to external tools. Popular servers: github, postgres, linear, notion, sentry, slack. Configure in ~/.claude/mcp.json.
6. Hooks
PreToolUse, PostToolUse, UserPromptSubmit, Stop, Notification, SubagentStop. Define in settings.json. Harness-enforced (not Claude-enforced) — safe to rely on for security.
7. Sub-Agents
Task tool delegates work to sub-agents. Built-in: general-purpose, code-reviewer, Plan, Explore. Custom in ~/.claude/agents/*.md.
8. Use Cases
- Repo onboarding
- Migration (Express → Next.js)
- Test coverage improvement
- Production bug fix (with Sentry MCP)
- KVKK audit (custom sub-agent)
- Schema migration (with Postgres MCP)
- Multi-service refactor
- i18n completion
9. KVKK Compliance
For sensitive code/IP, use Claude Team or Enterprise. Anthropic EU region (Frankfurt) for data residency. Hooks can enforce TC kimlik no / IBAN / credit card regex blocking.
10. Conclusion
Claude Code is the gold standard for agentic coding. CLAUDE.md + hooks + sub-agents combination automates 40-60% of senior engineer tasks. Strong Turkish fluency + KVKK alignment make it a top choice for Turkish developers.
Consulting Pathways
Consulting pages closest to this article
For the most logical next step after this article, you can review the most relevant solution, role, and industry landing pages here.
AI Agents and Workflow Automation
Move beyond single-step chatbots to AI workflows orchestrated with tools, rules and human approval.
Enterprise RAG Systems Development
Production-grade RAG systems that provide grounded, secure and auditable access to internal knowledge.
Enterprise AI Architecture Consulting for CTOs
Technical leadership consulting to move AI initiatives from isolated PoCs into secure, scalable and production-ready architecture.