#Claude Code
695 related articles
Claude Code Auto Mode Enabled by Default: A Guide to Keeping, Tuning, or Disabling It
Claude Code now defaults to auto mode with its safety classifier moved server-side. We break down the impact and give practical advice on keeping, tuning, or disabling auto mode, plus the v2.1.302 headless configuration trap.
Claude Code Cost Varies 11.9x: A Guide to Saving Money with Effort Subagents
Claude Code's official cost table reveals Effort thinking budget tiers: the same model on the same task can cost 11.9x more. This article breaks down each tier's cost index, how to assign work to subagents individually, and three default-setting pitfalls, plus three practical commands to finely control AI coding expenses.
Claude Code Subagents in Practice: Offload the Grunt Work to AI Clones and Keep the Main Session Light
A deep dive into Claude Code subagents: built-in Explore, Plan, and General agents, custom Markdown resumes with tool whitelists, parallel module investigation, and git worktree isolation—keeping the main session focused on judgment and decisions.
The Complete Guide to Claude Code: From Prompts to Subagents and Skills
From installation and login, CLAUDE.md configuration, Plan Mode refactoring, to MCP extensions, subagent creation, and skill reuse—a complete walkthrough of Claude Code's workflow and practical tips, plus a low-cost local model practice plan.
Dissecting Cursor's 10K-Star Plugin Repository: Install into Claude Code with One Command
Dissecting Cursor's official plugin repository: only one required field, auto-loaded components, and one command to install a Cursor plugin into Claude Code. Skill format works across clients, the official validator rejects extra fields, and we compare with DeepSeek's community-driven ecosystem.
Close Your Laptop and Let Claude Keep Coding: How AI Agents Take Over Tasks
What's the essential difference between chatbots and AI agents? This article, based on a YouTube video, explains Claude Code's agent loop: reading projects, editing files, running tests, auto-fixing, and why important operations still need human confirmation.
Claude Code Idle Auto-Compaction: Mechanism, Cache Math, and How to Disable It
Claude Code 2.1.299 auto-compacts sessions after 200K tokens and ~54 minutes idle. Learn the trigger mechanism, cache cost logic, and how to disable it.
Claude Code in Practice: AI Programming from Installation to Full-Stack Project Development
Based on Mosh Hamadani's Claude Code course, this guide covers installation and setup, prompt techniques, Plan Mode, context window management, MCP extensions, and engineering methodology for bug fixing and refactoring—helping you build production-grade full-stack apps from scratch.
Hands-On with Anthropic's Official claude-code-setup Plugin: Tailored Hooks and Skills Even for a Zero-Config Repo
Hands-on test of Anthropic's official claude-code-setup plugin: on a nearly zero-config Chrome extension repo, it recommends tailored Hooks, Skills, Subagents, and MCP servers based on file size, version-sync scripts, and PR history—entirely read-only with no side effects.
Finally Understanding What MCP Actually Does in Claude Code
A developer's take on what MCP actually does in Claude Code: it doesn't make Claude smarter, it lets it reach GitHub, databases, browsers, and other external systems. Plus the common myth that more MCP servers is better.
Build an AI Coworker with Claude Code That Works While You Sleep
A Claude API tutorial author uses Claude Code routines to build a scheduled cloud AI coworker: monitors 12 official doc pages each morning, generates Japanese summaries to Slack, and auto-drafts articles when the book manuscript is affected. This article breaks down the architecture, prompt design, pitfalls from the first two runs, and a cost comparison.
Claude Code Desktop 256KB Preview Flaw: SSH Keys Could Be Stolen
Noma Labs disclosed a high-severity flaw in Claude Code Desktop's preview panel: HTML over 256KB bypassed CSP, read SSH keys and AWS credentials, and exfiltrated them. Learn the chain, the patch bypass, and how to check yourself.
How to Check Claude Code, Codex, and Cursor Usage and Cost: A Complete Guide
Learn how to check token usage and real costs for Claude Code, Codex, and Cursor. Analyze local log files, compare plans vs API pricing, with test data and tips.
I Tested 28 Claude Code Skills — Here Are the 5 I Kept
After testing 28 Claude Code skills, one developer kept only 5: ponytail, caveman, graphify, grill-me, and ui-ux-pro-max. The core filter: do the tokens each skill loads justify the change in outcomes?
Claude Code v2.1.300 Update: OpenTelem…
Claude Code v2.1.300 Update: OpenTelemetry Enhancements and Multiple Core Fixes
A detailed look at Claude Code v2.1.300: new OpenTelemetry cache token split and GenAI semantic attributes, expanded mod system virtualText and agent.spawn hooks, plus fixes for cloud sessions, the VS Code extension, and remote control.
Anthropic's Official Counter-Example: Replacing "Fix the Login Bug" with a Six-Field Task Brief
Anthropic's docs use "fix the login bug" as a bad example. We break down a six-field task brief template (goal, scope, out of scope, constraints, done-check, report) that fuses guidance from Anthropic, OpenAI, and GitHub to make coding agents guess less and produce mergeable diffs.
Claude Code v2.1.299 Update: Idle Compaction and 50+ Fixes Explained
Claude Code v2.1.299 released with idle compaction, CLAUDE_CODE_DISABLE_FILE_MENTIONS, and 50+ fixes covering MCP connections, enterprise gateways, terminal rendering, and memory leaks.
Claude Code Installation, Configuration, and Anti-Ban Practical Guide
A detailed guide to installing and configuring Claude Code in China, covering NPM installation steps, Node.js setup, and using CC Switch to bypass account login and avoid bans, plus model selection advice.
Why Does Claude Code Load None of Its 170 MCP Tools?
Claude Code uses tool search to load only MCP tool names at startup instead of full manuals, cutting a 55,000-token tool overhead by over 85%. This article examines how it works, its real-world results, and its limits.
Measured Claude Code /compact Costs: $0.06 with Warm Cache, 9x Pricier Cold
Measured Claude Code /compact real cost: only $0.06 with a warm 136k-token cache, 9x pricier ($0.54) once cache expires, /clear costs zero API requests. Breakdown of how cache breakpoints can spike compaction costs 7x, with reproducible scripts.