<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>ShipWithAI Guides</title><link>https://shipwithai.online/guides/</link><description>Step-by-step guides for building with AI coding tools.</description><language>en</language><atom:link href="https://shipwithai.online/guides/feed.xml" rel="self" type="application/rss+xml"/><lastBuildDate>Fri, 16 Oct 2026 12:00:00 +0000</lastBuildDate><item><title>How to make a FiveM script with AI (step by step)</title><link>https://shipwithai.online/guides/fivem-script-with-ai/</link><guid>https://shipwithai.online/guides/fivem-script-with-ai/</guid><description>Use an AI coding tool to write, check and test a FiveM Lua script: fxmanifest, client and server events, and why the server must never trust the client.</description></item><item><title>How to choose an AI tool for FiveM</title><link>https://shipwithai.online/guides/choose-ai-tool-for-fivem/</link><guid>https://shipwithai.online/guides/choose-ai-tool-for-fivem/</guid><description>Six things to check before you pick an AI coding tool for FiveM Lua: project context, local linting, doc grounding, server checks, cost and code privacy.</description></item><item><title>How to make a RedM script with AI</title><link>https://shipwithai.online/guides/redm-script-with-ai/</link><guid>https://shipwithai.online/guides/redm-script-with-ai/</guid><description>Write a small RedM Lua script with an AI coding tool: the rdr3 manifest, client and server events, input checks and testing before you load it.</description></item><item><title>Claude Code, Codex or Cursor: which fits how you work?</title><link>https://shipwithai.online/guides/claude-code-codex-cursor/</link><guid>https://shipwithai.online/guides/claude-code-codex-cursor/</guid><description>Claude Code, Codex and Cursor side by side: plans, models, where each runs, data settings and limits, from each vendor&#x27;s own pages. No winner picked.</description></item><item><title>How to use AI to build a Godot 4.5+ feature, step by step</title><link>https://shipwithai.online/guides/ai-game-godot/</link><guid>https://shipwithai.online/guides/ai-game-godot/</guid><description>Build a dash move in Godot 4 with an AI coding tool: catch Godot 3 syntax like yield and export, check it headless, and test it before you open the editor.</description></item><item><title>Vibe coding for beginners: build your first app safely</title><link>https://shipwithai.online/guides/vibe-coding-for-beginners/</link><guid>https://shipwithai.online/guides/vibe-coding-for-beginners/</guid><description>New to AI app building? What vibe coding is, a safe first project, how to review AI code, use git, keep secrets out, test, and know when to stop and learn.</description></item><item><title>n8n, CrewAI or LangGraph: which one fits how you build?</title><link>https://shipwithai.online/guides/n8n-crewai-langgraph/</link><guid>https://shipwithai.online/guides/n8n-crewai-langgraph/</guid><description>n8n, CrewAI and LangGraph from their own docs and repos: how you build, what each is for, and each license as written. No winner picked.</description></item><item><title>How to stop AI coding tools breaking your code</title><link>https://shipwithai.online/guides/prompts-stop-ai-breaking-code/</link><guid>https://shipwithai.online/guides/prompts-stop-ai-breaking-code/</guid><description>Habits and copy-paste prompts that make AI code changes easier to check and undo: commit first, plan, set rules, run tests, read the diff, then roll back.</description></item><item><title>How to write an AGENTS.md (and CLAUDE.md) file for your AI coding tool</title><link>https://shipwithai.online/guides/agents-md-claude-md/</link><guid>https://shipwithai.online/guides/agents-md-claude-md/</guid><description>What AGENTS.md, CLAUDE.md, Cursor rules and Copilot instructions are, where each file goes and what takes priority, plus a sample AGENTS.md to copy.</description></item><item><title>Do AI coding tools train on your code? Settings by plan</title><link>https://shipwithai.online/guides/ai-coding-tools-train-on-your-code/</link><guid>https://shipwithai.online/guides/ai-coding-tools-train-on-your-code/</guid><description>What GitHub Copilot, Cursor, Claude Code, OpenAI Codex, Gemini Code Assist, Windsurf and JetBrains AI say about training on your code, by plan.</description></item><item><title>How to set up MCP servers in Claude Code, Cursor and Codex</title><link>https://shipwithai.online/guides/mcp-servers-claude-code-cursor-codex/</link><guid>https://shipwithai.online/guides/mcp-servers-claude-code-cursor-codex/</guid><description>Set up MCP servers in Claude Code, Cursor, Codex and Copilot in VS Code: config files, commands, project or global scope, connection checks, safety.</description></item><item><title>Keep secrets safe with AI coding tools: ignore files and permissions</title><link>https://shipwithai.online/guides/keep-secrets-safe-ai-coding-tools/</link><guid>https://shipwithai.online/guides/keep-secrets-safe-ai-coding-tools/</guid><description>How .cursorignore, Claude Code deny rules, Codex sandbox modes and Copilot content exclusion work, what each vendor says they can&#x27;t block, and key habits.</description></item><item><title>How to write Cursor rules: project rules and .mdc files, with examples</title><link>https://shipwithai.online/guides/cursor-rules-mdc/</link><guid>https://shipwithai.online/guides/cursor-rules-mdc/</guid><description>Write Cursor project rules in .cursor/rules: the .mdc format, alwaysApply, description and globs, the four rule types, glob patterns and what to avoid.</description></item><item><title>How to set up Claude Code hooks, with examples</title><link>https://shipwithai.online/guides/claude-code-hooks/</link><guid>https://shipwithai.online/guides/claude-code-hooks/</guid><description>Set up Claude Code hooks: where they go, PreToolUse and PostToolUse, exit codes 0 and 2, a Prettier example, a block-rm example, and the security rules.</description></item><item><title>How to write GitHub Copilot custom instructions: copilot-instructions.md and path-specific files</title><link>https://shipwithai.online/guides/copilot-custom-instructions/</link><guid>https://shipwithai.online/guides/copilot-custom-instructions/</guid><description>Write GitHub Copilot custom instructions: copilot-instructions.md, path-specific .instructions.md files with applyTo, AGENTS.md, and where each one works.</description></item><item><title>How to install Claude Code and run your first task</title><link>https://shipwithai.online/guides/install-claude-code/</link><guid>https://shipwithai.online/guides/install-claude-code/</guid><description>Install Claude Code on macOS, Linux, WSL or Windows, check it works, sign in, and run a first task safely. Commands quoted from Anthropic&#x27;s docs.</description></item><item><title>How to install the Codex CLI and run your first task</title><link>https://shipwithai.online/guides/install-codex-cli/</link><guid>https://shipwithai.online/guides/install-codex-cli/</guid><description>Install OpenAI&#x27;s Codex CLI on macOS, Linux or Windows, sign in, pick a safe sandbox mode, and run a first task. Commands quoted from OpenAI&#x27;s docs.</description></item><item><title>How to install Cursor and run your first task</title><link>https://shipwithai.online/guides/install-cursor/</link><guid>https://shipwithai.online/guides/install-cursor/</guid><description>Install the Cursor editor on macOS, Windows or Linux, sign in, turn on Privacy Mode, and make a first small change safely. Steps quoted from Cursor&#x27;s docs.</description></item><item><title>How to set up GitHub Copilot in VS Code and run your first task</title><link>https://shipwithai.online/guides/setup-copilot-vscode/</link><guid>https://shipwithai.online/guides/setup-copilot-vscode/</guid><description>Install VS Code, sign in to GitHub Copilot, turn off training on your data, and run a first agent task safely. Steps quoted from VS Code and GitHub docs.</description></item><item><title>How to review AI-written code before you keep it</title><link>https://shipwithai.online/guides/review-ai-code/</link><guid>https://shipwithai.online/guides/review-ai-code/</guid><description>A beginner checklist for checking an AI agent&#x27;s changes: diffs, tests, secrets, risky files and each tool&#x27;s own undo. Quoted from vendor and OWASP docs.</description></item><item><title>Git basics for AI coders: save points, undo and branches</title><link>https://shipwithai.online/guides/git-basics-ai-coders/</link><guid>https://shipwithai.online/guides/git-basics-ai-coders/</guid><description>A plain beginner guide to Git for people coding with AI: commit save points, check changes, undo safely, and try ideas on branches. Quoted from Git docs.</description></item><item><title>AI coding agent deleted your files? Recover them, then stop it happening</title><link>https://shipwithai.online/guides/ai-agent-deleted-files/</link><guid>https://shipwithai.online/guides/ai-agent-deleted-files/</guid><description>Claude Code, Cursor or Codex deleted or overwrote files? What to try first to get them back, why deny rules alone don&#x27;t stop it, and the settings that do.</description></item><item><title>AI invented a FiveM native? How to check it</title><link>https://shipwithai.online/guides/fivem-ai-hallucinated-natives/</link><guid>https://shipwithai.online/guides/fivem-ai-hallucinated-natives/</guid><description>An AI wrote a FiveM native or QBCore export that does not exist. Here is how to check any name against the official docs in 2 minutes.</description></item><item><title>How to give an AI coding agent a clear task: scope, context and done-checks</title><link>https://shipwithai.online/guides/clear-agent-tasks/</link><guid>https://shipwithai.online/guides/clear-agent-tasks/</guid><description>Write agent tasks with a goal, context, constraints and a done-check. Covers @ and # file mentions, examples, tests and evidence, quoted from vendor docs.</description></item><item><title>AI agent permissions for beginners: approval modes, allow/deny rules and sandboxes</title><link>https://shipwithai.online/guides/agent-permissions-approvals/</link><guid>https://shipwithai.online/guides/agent-permissions-approvals/</guid><description>How approval modes, allow/deny rules and sandboxes work in Claude Code, Codex, Cursor and VS Code Copilot, quoted from vendor docs. Checked Oct 2026.</description></item><item><title>Before you ship: lock the database, back it up, and log what happens</title><link>https://shipwithai.online/guides/database-exposure-backups-logging/</link><guid>https://shipwithai.online/guides/database-exposure-backups-logging/</guid><description>A pre-launch checklist for AI-built apps: RLS, IP allowlists, Postgres trust auth, Firebase rules, tested restores and safe logging, quoted from docs.</description></item><item><title>Deploy safely with a way to roll back</title><link>https://shipwithai.online/guides/safe-deploys-rollback/</link><guid>https://shipwithai.online/guides/safe-deploys-rollback/</guid><description>A beginner checklist for AI-built apps: preview first, know your rollback button, git revert vs reset, migrations and gradual rollout, quoted from docs.</description></item><item><title>Run things at the same time safely</title><link>https://shipwithai.online/guides/concurrency-safety/</link><guid>https://shipwithai.online/guides/concurrency-safety/</guid><description>A beginner checklist for AI-built apps and game scripts: race conditions, transactions, row locks, retries and double submits, quoted from docs.</description></item><item><title>Add user login safely to an AI-built app</title><link>https://shipwithai.online/guides/auth-login-safely/</link><guid>https://shipwithai.online/guides/auth-login-safely/</guid><description>A beginner checklist for adding login to AI-built apps: password storage, generic errors, login limits, session cookies, MFA and checks, quoted from docs.</description></item><item><title>Keep an AI-built app safe from prompt injection</title><link>https://shipwithai.online/guides/prompt-injection-safety/</link><guid>https://shipwithai.online/guides/prompt-injection-safety/</guid><description>A beginner checklist for prompt injection in AI-built apps: untrusted text, limited tools, approvals, no secrets in prompts and logging, quoted from docs.</description></item><item><title>Claude Code /compact vs /clear: which one to use</title><link>https://shipwithai.online/guides/compact-vs-clear/</link><guid>https://shipwithai.online/guides/compact-vs-clear/</guid><description>What /compact and /clear do in Claude Code, what each keeps, how auto-compact works, and a simple way to pick one. From Anthropic&#x27;s own docs.</description></item><item><title>How to get out of an AI bug loop</title><link>https://shipwithai.online/guides/ai-bug-loop/</link><guid>https://shipwithai.online/guides/ai-bug-loop/</guid><description>Your AI coding agent keeps fixing and re-breaking the same bug. How to stop, roll back, start fresh and give it a test it can&#x27;t fake. From vendor docs.</description></item><item><title>The AI told me to install a package. Is it real?</title><link>https://shipwithai.online/guides/slopsquatting/</link><guid>https://shipwithai.online/guides/slopsquatting/</guid><description>AI coding tools sometimes suggest packages that don&#x27;t exist, and attackers can register those names. How to check a package before you install it.</description></item><item><title>Cursor Ask vs Agent vs Plan mode: which one to use</title><link>https://shipwithai.online/guides/cursor-ask-agent-plan-modes/</link><guid>https://shipwithai.online/guides/cursor-ask-agent-plan-modes/</guid><description>What Cursor&#x27;s Ask, Agent, Plan and Debug modes do, which ones edit your files, how to switch, and a simple way to pick one. From Cursor&#x27;s own docs.</description></item><item><title>How to make custom slash commands (skills) in Claude Code</title><link>https://shipwithai.online/guides/claude-code-skills-slash-commands/</link><guid>https://shipwithai.online/guides/claude-code-skills-slash-commands/</guid><description>Make your own /commands in Claude Code with a SKILL.md file: where to save it, the key frontmatter fields, arguments, and safety. From Anthropic&#x27;s docs.</description></item><item><title>How to create a Claude Code subagent</title><link>https://shipwithai.online/guides/claude-code-subagents/</link><guid>https://shipwithai.online/guides/claude-code-subagents/</guid><description>What Claude Code subagents are, when to use one, how to write the file, where to save it and how to call it. From Anthropic&#x27;s own docs.</description></item><item><title>Claude Code in VS Code vs the terminal: which should you start with?</title><link>https://shipwithai.online/guides/claude-code-vscode-vs-terminal/</link><guid>https://shipwithai.online/guides/claude-code-vscode-vs-terminal/</guid><description>What Anthropic&#x27;s docs say about the Claude Code VS Code extension and the terminal CLI: setup, what each can do, and how to switch between them.</description></item><item><title>Your first n8n workflow with an AI Agent node: chat, model, memory and tools</title><link>https://shipwithai.online/guides/n8n-first-ai-agent-workflow/</link><guid>https://shipwithai.online/guides/n8n-first-ai-agent-workflow/</guid><description>Build a first n8n chat workflow with the AI Agent node: Chat Trigger, a chat model, memory, tools and human review, quoted from n8n&#x27;s own docs.</description></item><item><title>Upgrade a Godot 3 project to Godot 4: the converter first, then AI for the rest</title><link>https://shipwithai.online/guides/godot-3-to-4-upgrade-ai/</link><guid>https://shipwithai.online/guides/godot-3-to-4-upgrade-ai/</guid><description>Back up, run Godot&#x27;s project upgrade tool, then use an AI tool to fix what it can&#x27;t convert. Quoted from the official Godot docs, checked Oct 2026.</description></item><item><title>Check Supabase RLS before you launch an AI-built app</title><link>https://shipwithai.online/guides/supabase-rls-before-launch/</link><guid>https://shipwithai.online/guides/supabase-rls-before-launch/</guid><description>A docs-checked checklist for Row Level Security on Supabase before launch: enable RLS, know your keys, run the Security Advisor, and avoid open policies.</description></item><item><title>FiveM server event security: source checks, fake payouts and heavy loops</title><link>https://shipwithai.online/guides/fivem-server-event-security/</link><guid>https://shipwithai.online/guides/fivem-server-event-security/</guid><description>Docs-checked FiveM guide: validate source on net events, keep purchases server-side, review AI-written resources, and find wasteful per-frame loops.</description></item></channel></rss>
