All Articles

Data Science & Software Writing by Ahmet Bilgiç

Back to Homepage Articles

Claude Code Permissions Guide

allow/ask/deny rules, Bash/Read/Edit/WebFetch/MCP rule syntax, permission modes, wildcard pitfalls, extending with hooks and managed settings. (Turkish)

Claude Code Sandbox Guide

Running the Bash tool with filesystem and network isolation: /sandbox setup, auto-allow mode, allowWrite/denyRead, credential protection and limitations. (Turkish)

Claude Code Auto Mode Guide

The classifier-based permission mode: decision order, what's blocked/allowed by default, autoMode.environment, allow/soft_deny/hard_deny rules and limitations. (Turkish)

Claude Code Plugin & Marketplace Guide

Creating plugins, testing with --plugin-dir, plugin.json and directory structure, adding marketplaces, the official/community marketplaces and extraKnownMarketplaces. (Turkish)

Claude Code Checkpoint and /rewind Guide

Automatic checkpoints, the /rewind menu, restoring code/conversation, summarizing, limitations, and how it compares to /branch. (Turkish)

Claude Code on the Web Guide

--cloud, --teleport, cloud environments, auto-fix pull requests and Routines: the complete guide to running Claude Code in the cloud. (Turkish)

Claude Code Cross-Session Messaging Guide

SendMessage and ListAgents for messaging between sessions: /list-agents, delivery rules, crossSessionInbound, isolatePeerMachines and cross-machine messaging. (Turkish)

Claude Code Review Guide

The /code-review command, effort levels, --fix/--comment flags, cloud-based ultrareview and automated PR review on GitHub. (Turkish)

Claude Code Workflows Guide

The agent(), pipeline(), parallel() script API, the ultracode keyword, the /workflows command, the approval flow, limits and saving runs for multi-agent orchestration. (Turkish)

Claude Opus 5 in Claude Code

Which accounts default to it, the /model and /effort commands, the thinking behavior change, fast mode and what's new in v2.1.219. (Turkish)

GPT-5.6: Sol, Terra and Luna

OpenAI's new model family: how the three models differ, the 1.05M-token context window, API pricing, reasoning effort/mode settings and which one to pick. (Turkish)

Claude Code GitHub Actions Guide

/install-github-app setup, workflow YAML, claude_args, triggers, Bedrock/Vertex support and automated PR review for Claude Code + GitHub Actions. (Turkish)

Claude Code Subagents Guide

Built-in agents (Explore, Plan), frontmatter fields, model selection, tool restrictions, hooks, and chaining workflows for subagents in Claude Code. (Turkish)

Claude Code MCP Guide

Adding servers (claude mcp add), scopes, .mcp.json, OAuth authentication, resources and security for the Model Context Protocol in Claude Code. (Turkish)

Claude Code Hooks Guide

PreToolUse, PostToolUse, SessionStart and Notification events, JSON schema, exit codes and real examples for event-driven automation in Claude Code. (Turkish)

Claude Code: Agent Skills & Slash Commands

SKILL.md structure, frontmatter fields, arguments, dynamic context injection and running skills in a subagent — a guide to writing custom skills in Claude Code. (Turkish)

Claude Code: CLAUDE.md & Auto Memory

CLAUDE.md file locations, load order, @import syntax, .claude/rules/, and Auto Memory — a guide to persistent project memory in Claude Code. (Turkish)

Claude Code CLI & Automation

Print mode, output formats, permission modes and --allowedTools — using Claude Code from scripts and CI. (Turkish)

Getting Started with Claude Code

Komut İstemcisinde Claude Code Kullanmak İçin Yapılması Gerekenler — installing, logging in and first steps with Claude Code in the terminal. (Turkish)

Outliers in Data Science

Aykırı Değerler: Tanımı, Tespiti ve Yönetim Stratejileri — detecting and managing outliers with IQR, Z-score, Mahalanobis and LOF. (Turkish)

Handling Missing Data

Veri Biliminde Eksik Veriler ile Nasıl Başa Çıkarız? — from dropping rows to model-based imputation, with Python examples. (Turkish)

Imbalanced Datasets

Dengesiz Veri Setleri ile Nasıl Başa Çıkılır? — SMOTE, resampling and the right metrics, through a credit card fraud case study. (Turkish)