Articles: Claude Code
Claude Code compact: context checks and continuation
Use /context and focused /compact prompts, then verify the saved work and source details before continuing.
Claude Code sandbox: files, network and strict mode
Claude Code sandbox limits for files and networks, strict mode, retries and the tools outside its boundary.
CLAUDE.md vs AGENTS.md: which file each coding agent reads
Claude Code now reads AGENTS.md when no CLAUDE.md exists. Here is what Codex, Copilot, Cursor and Gemini CLI read.
Where Claude Code stores sessions, settings and memory
Claude Code keeps sessions in ~/.claude/projects/ for 30 days, settings in settings.json files and memory per repo.
Claude Desktop vs Claude Code: which workspace fits the job?
Claude Desktop and Claude Code share a model family but organise work differently. Here is the practical split.
Example CLAUDE.md files for Bun, Python and React
Adapt real CLAUDE.md files from Anthropic's Bun, Python and React projects, with sourced commands and pitfalls.
Serve Claude Code as MCP: Setup and Limits
Run Claude Code as a local stdio MCP server, configure the client and keep approvals at the client boundary.
Global CLAUDE.md: personal instructions across projects
Global CLAUDE.md: file location, personal preferences, project scope and checks for loaded instructions.
Claude skills marketplace: where to find and install skills
Where the Claude skills marketplace lives: Claude Code plugin marketplaces, /plugin install and claude.ai skills.
Claude Code skills: SKILL.md, locations, frontmatter and install
Claude Code skills are SKILL.md folders Claude loads when relevant or you run as /skill-name. Where they go and how.
OpenClaw vs Claude Code: purpose, models, license, security
OpenClaw vs Claude Code: OpenClaw is a self-hosted MIT assistant in chat apps; Claude Code is Anthropic's coding agent.
Antigravity vs Claude Code: models, plans, free tier, MCP
Antigravity vs Claude Code: Google's agent app runs Gemini with a $0 tier; Claude Code runs Claude on paid Claude plans.
Get Shit Done for Claude Code: GSD Core install and commands
Get Shit Done is a community framework that runs Claude Code in a plan, execute, verify loop, now continued as GSD Core.
How much is Claude Code? Pro, Max, Team and API prices
How much is Claude Code? It comes with paid Claude plans: Pro $20/mo, Max $100 or $200, Team $20-$100 per seat.
Claude PDF skill: what it does and how to install it
The Claude PDF skill is Anthropic's skill for reading, creating, merging, splitting, filling and OCRing PDFs.
CLAUDE.md best practices: locations, imports and examples
CLAUDE.md best practices: run /init, keep it under about 200 lines, and only include what Claude can't infer from code.
Claude Code vs Codex: plans, config, MCP and sandboxes
Claude Code vs Codex: Claude Code needs a paid Claude or Console account; Codex is in every ChatGPT plan, even Free.
Claude vs Claude Code: chat apps vs agentic coding tool
Claude vs Claude Code: Claude is the chat app for writing and research; Claude Code edits and runs code in your repo.
Are MCP servers free? Protocol, licenses and service costs
The protocol is open and most MCP servers are open-source software you can run at no charge.
Claude Code hooks: events, matchers, JSON and exit codes
Claude Code hooks run commands, HTTP calls or prompts at events like PreToolUse and Stop, set in .claude/settings.json.
Claude Code features: surfaces, subagents, skills, hooks, MCP
Claude Code features: terminal, IDE, desktop, web and mobile surfaces, plus CLAUDE.md, skills, subagents, hooks and MCP.
Best MCP servers for Claude Code, grouped by job
Best MCP servers for Claude Code: official ones like GitHub, Playwright and Sentry, each added with claude mcp add.
Claude Code MCP: add, configure and manage servers
Claude Code MCP: add a server with claude mcp add --transport http <name> <url>, then check it with /mcp.
claude --dangerously-skip-permissions: risks and safer modes
claude --dangerously-skip-permissions starts Claude Code in bypassPermissions mode with no prompts.