#Claude Code
675 related articles
17 Claude Code Tips from Real-World Practice: Save Tokens, Avoid CLAUDE.md Pitfalls, Build Efficient Workflows
17 Claude Code tips dissected: shrink CLAUDE.md to save tokens, use /context to diagnose context rot, protect key info with /compact, subagents vs. agent teams, and how to verify UltraThink, Context 7, and Chrome DevTools.
Claude Code Hooks in Practice: Auto-Format Code and Block Sensitive Operations
Claude Code Hooks tutorial: use pre/post tool call events to auto-format code, block sensitive files, send desktop notifications, and configure loop patrols so your AI coding assistant manages itself.
Build and Deploy an App on Google Cloud with Claude Code in 30 Minutes
Google Cloud developer advocate Ivan Nardini demonstrates building and deploying a feedback app with Claude Code and Google Cloud in 30 minutes, covering plan mode, MCP servers, Skills, parallel sub-agents, and security review — showing how AI coding agents span the enterprise software lifecycle.
Running Claude Code with Local Ollama: A Complete Hands-On Guide
A step-by-step guide to connecting Claude Code to local Ollama models across Windows, Linux, and macOS, covering both the ollama launch and environment variable methods, plus 8 GB RAM test findings and a troubleshooting checklist.
An Unattended Claude Loop Burned 13% of My Quota Overnight — Here Are the Three Safeguards I Added
An unattended Claude loop ran all night and burned 13% of plan quota. Here's how Claude limits are shared with the chat app, how to read /usage, and three safeguards for automated loops.
Claude Code v2.1.304 Update Explained: Auto Mode and Permission Security Upgrades
Claude Code v2.1.304 explained: new /auto-mode-setup guidance, decider and isReadOnly permission fields for tool.check, plus dozens of fixes for multi-agent, plugin hot-reload, and context management.
Claude Code Installation and Advanced Guide: Bypass Account Ban Risks, Master Permissions, MCP, and Skills
Claude Code installation guide for China: NPM method to bypass network limits, cc-switch to avoid account bans, plus permission modes, MCP, Skills, and model tips.
Claugravity: One-Click Migration of Claude Code Plugins to Antigravity
Claugravity is an open-source tool that converts Claude Code plugins into Google Antigravity's SKILL.md format, offering static security scanning, conflict handling, and a local web dashboard, with support for Windows, macOS, and Linux.
7 Things AI Coding Agents Can't See, and the Command That Shows Each One
AI coding agents only see repo files and your typed text; CI logs, issue threads, PR inline comments, production schema, runtime logs, feature flag state, and the browser console are all beyond their view. Here are seven commands that turn each hidden state into agent-readable text, plus how .context/ and CLAUDE.md keep context across sessions.
Claude Code Usage Stats Trap: Why One API Response Gets Counted Multiple Times
Claude Code transcripts can split one API response across multiple lines, each carrying the full usage object. Naive summation overestimates token usage by nearly 2x. This article explains Anthropic's message.id deduplication rule, the output token placeholder trap, and Node and jq scripts for comparison.
Enterprise-Level Vibe Coding in Practice: Claude Code + Codex Engineering Development
A Bilibili hands-on course breakdown: using Claude Code and Codex to go from Vibe Coding to Super Power engineering-driven development of enterprise e-commerce and an OpenRouter-like aggregation platform, comparing tool capabilities and model selection while revealing the real boundaries of AI programming.
Claude Code Cost Breakdown: 43% Spent on System Prompts
A developer's Claude Code token log: 2,772 daily calls, fixed context (system prompt + tools) at 43.3% of cost, model replies only 25.5%. Shortening sessions saves more than trimming tool output.
Running Multiple Claude Code Agents in Parallel: The Hidden Pitfalls of Worktrees
Git worktrees only isolate the filesystem. Parallel Claude Code agents still collide on ports, databases, env vars, and credentials. Here's what to do about it.
Lost Your Claude Code Session After Accidentally Closing the Terminal? This Open-Source Plugin Can Find It
termstash is an open-source plugin for Claude Code that recovers sessions lost after accidentally closing the terminal. Use npx termstash last without installation, or install it as a plugin and tell Claude "I lost my session."
Claude Code Preview Window Vulnerability: The Attack Surface Beyond Agent Security
Claude Code Desktop's preview window had a high-severity flaw: malicious HTML could steal SSH keys and cloud credentials. Fixed now, but the agent security boundary lesson matters.
The Top 1% Way to Use Claude Code: Drop Plan Mode and Use Verifiable Criteria to Boost Quality
Top Claude Code users have dropped plan mode. This article breaks down the top 1%'s practices: verifiable criteria that boost quality 2-3x, /goal auto-loops, a lean CLAUDE.md as a table of contents, skills for error correction only, hooks instead of all-caps rules, sub-agent limits, and scheduled tasks for unattended work.
Nine Advanced Claude Code Features in Practice: Subagents, Skills, Hooks, and Headless Mode Explained
Nine hands-on demos explain Claude Code subagents, skill locking, hooks automation, MCP integration, worktree parallelism, headless mode, checkpoint rollback, and session branching.
Linux Creator Uses Claude Code to Write Kernel? AI Programming Controversy and Practice
Linux creator Linus Torvalds was reported to use Claude Code to write kernel code, sparking AI programming controversy. This article reviews the event details, analyzes performance improvement data, and shares cloud development anti-ban tips, exploring the future trends of AI programming.
Claude Code Lead: Since Opus 4.5, 100% of My Code Is Written by AI
Anthropic Claude Code lead Boris reveals that since Opus 4.5, 100% of his code is written by Claude: 1,700 PRs, 400K lines added, 8B tokens. A deep dive into dynamic workflows, loops as an abstraction layer, Auto Mode, and the ROI logic of AI programming.
225 Billion Tokens Measured: The Truth About Claude Code Costs and Multi-Agent Workflows
A developer's 30-day Claude Code ledger: 225.1B tokens, 97.9% cache re-reads, 1,029 merged PRs, and 59%-80% of subagents inheriting costly Opus. Cost structure, span of control, and configuration pitfalls.