TL;DR Your Pro tip 1: before adding something to Here are some parts of my Also see:<\/p>\n For project-specific instructions, put them in the project-level The highest-signal content in your project Also see:<\/p>\n For language-specific or per-file rules<\/a>, put them in For instance, The full rules I have:<\/p>\n Claude Code provides a built-in method to modify the system prompt to change how Claude responds: Output styles<\/a>. For instance, mine forces ASD-STE100 Simplified Technical English, so every reply reads like an aircraft maintenance manual: short sentences, simple words, active voice.<\/p>\n Why an output style instead of the global One gotcha: output styles apply to the main conversation only \u2014 a subagent runs its own system prompt, so your style doesn't change how subagents respond.<\/p>\n There are some useful configurations<\/a> you could set in your Highlights:<\/p>\n I turned all this settings-tweaking into a skill: audit-claude-settings<\/a> fetches the latest settings and env-vars from official docs, diffs them against your actual config, and suggests changes tied to how you work.<\/p>\n The full settings I use:<\/p>\n If you're not using a sandbox or devcontainer for Claude Code, you may want to block some evil<\/strong> commands in your However, Though, Claude Code can still write a one-time script to read sensitive data<\/strong> and bypass all of the above defenses. So the safest approach is using sandbox after all.<\/p>\n Claude Code Plugins<\/a> are simply a way to package skills, commands, agents, hooks, and MCP servers. Distributing them as a plugin has the following advantages:<\/p>\n To install a plugin, you need to add a marketplace first. A marketplace is usually just a GitHub repo. Think of it as a namespace.<\/p>\n Recommended:<\/p>\n Skills<\/a> can contain executable scripts and hooks, not just Markdown. Use with caution!<\/strong> When in doubt, have your agent review them first.<\/p>\n Here are skills I've used, mostly installed per project when needed:<\/p>\n Recommended:<\/p>\n You can find more skills on skills.sh<\/a>.<\/p>\n You probably don't need any MCP servers<\/a> if you can do the same thing with CLI + skills.<\/p>\n No, just use the No, you should use the No, you should use the Yes, ironically. Other coding agents like Claude Code can use Codex via MCP, which is slightly more stable than directly invoking it with However, since OpenAI releases the official Claude Code plugin: codex-plugin-cc<\/a>, you should probably use that instead.<\/p>\n Both Claude Code<\/a> and Codex<\/a> support hooks. Hooks make Claude Code run specific commands on lifecycle events like Instead of reminding Claude Code to run the linter or tests in your prompts (and it still forgets sometimes), just write a I also wrote some Claude Code plugins that use hooks:<\/p>\n For example, hal-session-auto-rename<\/a>. Now that Claude can message your other Claude Code sessions<\/a> by name, a good session name actually matters. I found Claude Code already auto-names every session in the transcripts, so I just wired that up to a Use
\nMy configs, plugins, and skills for Claude Code:
\nhttps:\/\/github.com\/vinta\/hal-9000<\/a><\/p>\n<\/blockquote>\nCLAUDE.md<\/h2>\n
The Global CLAUDE.md<\/h3>\n
~\/.claude\/CLAUDE.md<\/code> should only contain:<\/p>\n\n
CLAUDE.md<\/code>, ask it, "Is this already covered in your system prompt?"<\/strong>
\nPro tip 2: try my refactor-claude-md<\/a> skill!<\/p>\nCLAUDE.md<\/code> I found useful:<\/p>\n### Use AskUserQuestion\n\nWhen you need input and the answer is a selection rather than a sentence (multiple-choice, yes\/no confirmations that gate next steps, picking from a list, choosing between approaches), ask with the AskUserQuestion<\/code> tool instead of plain text, so the user clicks an option instead of typing.\n\nWhen presenting approaches, put the summary in each option's label and the pros\/cons in its description. Plain text is fine when the answer is open-ended. This changes the format of questions, not whether to ask: never use it to ask permission for work you already have enough information to do.\n\n### Prefer online sources\n\nTraining data goes stale: library\/framework\/SDK APIs, config keys, CLI flags, cloud services, platform features, syntax, and versions change, and guessing has repeatedly cost debugging round-trips.\n\nInvoke the find-docs<\/code> skill BEFORE writing code or config that touches any of those, and BEFORE answering questions about them. Being about to write such code is trigger enough, even when no question was asked. Confidence is not an exemption, and neither is the library being well known. Answering from training data, or fetching a remembered docs URL instead of invoking the skill, does not satisfy this rule.\n\nIf the user provides URLs, WebFetch<\/code> each one as a primary source before searching further. Never skip user-provided URLs. For topics find-docs<\/code> covers poorly, WebFetch<\/code> the official docs instead of falling back to training data.\n\n### Auto-commit\n\nWhen you have completed the user's requested change, use the commit<\/code> skill to commit, always passing why the changes were made (e.g. \/commit add login endpoint for mobile app auth<\/code>); when no reason was stated, pass the request that prompted the changes instead \u2014 never an invented why.\n\nDon't batch unrelated changes into one commit.<\/code><\/pre>\n\n
The Project CLAUDE.md<\/h3>\n
CLAUDE.md<\/code>.<\/p>\nCLAUDE.md<\/code> (or any skill) is the Gotchas<\/strong> section. Build these from the failure points Claude Code actually runs into.<\/p>\n\n
Per File Type Rules<\/h2>\n
~\/.claude\/rules\/<\/code>, so Claude Code only loads them when editing those file types.<\/p>\n~\/.claude\/rules\/typescript-javascript.md<\/code>:<\/p>\n---\npaths:\n - \"**\/*.{ts,tsx}\"\n - \"**\/*.{js,jsx}\"\n - \"**\/package.json\"\n---\n\n# TypeScript \/ JavaScript\n\n- Pin exact dependency versions in package.json<\/code> \u2014 no ^<\/code> or ~<\/code> prefixes\n- Use node:<\/code> prefix for Node.js built-in modules (e.g., node:fs<\/code>, node:path<\/code>)\n- Write proper types\/interfaces instead of any<\/code> or casts like as any<\/code> \/ as unknown<\/code>\n - When a value is genuinely untypable, use unknown<\/code> and narrow it explicitly. any<\/code> is the last resort when no typed alternative exists\n- Prefer interface<\/code> over type<\/code> for object shapes (extendable, better error messages)\n- Avoid enums. Use union types (type Status = 'active' | 'inactive'<\/code>) or as const<\/code> objects\n- Mark properties and parameters readonly<\/code> when they should not be mutated\n- Do not add explicit return types. Let TypeScript infer them\n- Use the typescript<\/code> LSP tool for type-aware code navigation when grep's text matching would be ambiguous<\/code><\/pre>\n\n
Output Styles<\/h2>\n
---\nname: ASD-STE100\ndescription: Responses in ASD-STE100 Simplified Technical English\nkeep-coding-instructions: true\n---\n\nWrite all responses in ASD-STE100 Simplified Technical English.\n\n- Use the active voice.\n- Use the simple present, simple past, and simple future tenses. Do not use the -ing form of a verb.\n- Use a word with only one meaning. Use the simple word: \"start\", not \"commence\".\n- Write one instruction in each sentence.\n- Use a maximum of 20 words in an instruction. Use a maximum of 25 words in a description.\n- Write about one topic in each paragraph. Use a maximum of 6 sentences in a paragraph.<\/code><\/pre>\nCLAUDE.md<\/code>? They land in different places: an output style becomes part of the system prompt, and Claude Code periodically reminds the model to stick to it mid-conversation, while CLAUDE.md<\/code> gets injected as a user message, where it competes with all your other rules. Plus, you can switch styles in \/config<\/code> without touching your global rules.<\/p>\n\n
Configurations<\/h2>\n
Settings<\/h3>\n
~\/.claude\/settings.json<\/code>:<\/p>\n{\n \"$schema\": \"https:\/\/json.schemastore.org\/claude-code-settings.json\",\n \"env\": {\n \"CLAUDE_CODE_DISABLE_FEEDBACK_SURVEY\": \"1\",\n \"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS\": \"1\",\n \"CLAUDE_CODE_RETRY_WATCHDOG\": \"1\",\n \"DISABLE_ERROR_REPORTING\": \"1\",\n \"DISABLE_EXTRA_USAGE_COMMAND\": \"1\",\n \"DISABLE_FEEDBACK_COMMAND\": \"1\",\n \"DISABLE_UPGRADE_COMMAND\": \"1\"\n },\n \"permissions\": {\n \"allow\": [\"...\"],\n \"deny\": [\"...\"],\n \"ask\": [\"...\"],\n \"defaultMode\": \"auto\"\n },\n \"model\": \"opus[1m]\",\n \"effortLevel\": \"high\",\n \"advisorModel\": \"fable\",\n \"cleanupPeriodDays\": 365,\n \"includeGitInstructions\": false,\n \"showClearContextOnPlanAccept\": true,\n \"teammateMode\": \"auto\",\n \"voice\": { \"enabled\": true }\n}<\/code><\/pre>\n\n
"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1"<\/code>: Enable Agent Team<\/a> feature, a fancy way to consume a huge amount of tokens<\/li>\n"permissions.defaultMode": "auto"<\/code>: Use this to pretend it's safer than --dangerously-skip-permissions<\/code><\/li>\n"advisorModel": "fable"<\/code>: Use something faster like sonnet<\/code> as the main model, and let it consult<\/a> fable<\/code> when needed<\/li>\n"includeGitInstructions": false<\/code>: Remove built-in commit\/PR instructions and git status snapshot from the system prompt, since my commit<\/a> skill covers that<\/li>\n"cleanupPeriodDays": 365<\/code>: By default, your chat history (location: ~\/.claude\/projects\/<\/code>) will be deleted after 30 days<\/li>\n"voice": { "enabled": true }<\/code>: Enable Voice Dictation<\/a> feature. Code like a boss!<\/li>\n<\/ul>\n\n
Permissions<\/h3>\n
~\/.claude\/settings.json<\/code>:<\/p>\n{\n \"permissions\": {\n \"defaultMode\": \"auto\",\n \"deny\": [\n \"Read(~\/.aws\/**)\",\n \"Read(~\/.config\/**)\",\n \"Read(~\/.docker\/**)\",\n \"Read(~\/.dropbox\/**)\",\n \"Read(~\/.gnupg\/**)\",\n \"Read(~\/.gsutil\/**)\",\n \"Read(~\/.kube\/**)\",\n \"Read(~\/.npmrc)\",\n \"Read(~\/.orbstack\/**)\",\n \"Read(~\/.pypirc)\",\n \"Read(~\/.ssh\/**)\",\n \"Read(~\/*_history)\",\n \"Read(~\/**\/*credential*)\",\n \"Read(~\/Library\/**)\",\n \"Edit(~\/Library\/**)\",\n \"Read(~\/Dropbox\/**)\",\n \"Edit(~\/Dropbox\/**)\",\n \"Read(\/\/etc\/**)\",\n \"Edit(\/\/etc\/**)\",\n \"Bash(git -c *)\",\n \"Bash(git --config-env*)\",\n \"Bash(git --git-dir*)\",\n \"Bash(gh repo delete *)\",\n \"Bash(su *)\",\n \"Bash(sudo *)\",\n \"Bash(passwd *)\",\n \"Bash(env *)\",\n \"Bash(printenv *)\",\n \"Bash(history *)\",\n \"Bash(fc *)\",\n \"Bash(eval *)\",\n \"Bash(exec *)\",\n \"Bash(rsync *)\",\n \"Bash(sftp *)\",\n \"Bash(telnet *)\",\n \"Bash(socat *)\",\n \"Bash(nc *)\",\n \"Bash(ncat *)\",\n \"Bash(netcat *)\",\n \"Bash(nmap *)\",\n \"Bash(kill *)\",\n \"Bash(killall *)\",\n \"Bash(pkill *)\",\n \"Bash(chmod *)\",\n \"Bash(chown *)\",\n \"Bash(chflags *)\",\n \"Bash(xattr *)\",\n \"Bash(diskutil *)\",\n \"Bash(mkfs *)\",\n \"Bash(security *)\",\n \"Bash(defaults *)\",\n \"Bash(launchctl *)\",\n \"Bash(osascript *)\",\n \"Bash(dscl *)\",\n \"Bash(networksetup *)\",\n \"Bash(scutil *)\",\n \"Bash(systemsetup *)\",\n \"Bash(pmset *)\",\n \"Bash(crontab *)\"\n ],\n \"ask\": [\n \"Bash(curl *)\",\n \"Bash(wget *)\",\n \"Bash(open *)\",\n \"Bash(* install *)\",\n \"Bash(bun add *)\",\n \"Bash(yarn add *)\",\n \"Bash(pnpm add *)\",\n \"Bash(uv add *)\",\n \"Bash(git push *)\",\n \"Bash(git remote add *)\",\n \"Bash(git remote rename *)\",\n \"Bash(git remote remove *)\",\n \"Bash(git remote rm *)\",\n \"Bash(git remote set-url *)\",\n \"Bash(git config remote.*)\",\n \"Bash(git config * remote.*)\",\n \"Bash(gh repo create *)\",\n \"Bash(gh repo rename *)\",\n \"Bash(npx supabase db *)\"\n ]\n },\n \"hooks\": {\n \"PreToolUse\": [\n {\n \"matcher\": \"Bash\",\n \"hooks\": [\n {\n \"type\": \"command\",\n \"command\": \"python3 ~\/.claude\/hooks\/guard-bash-paths.py\"\n }\n ]\n }\n ]\n }\n}<\/code><\/pre>\n"deny": ["Read(~\/.aws\/**)", "Read(~\/.kube\/**)", ...]<\/code> alone is not enough, since Claude Code can still read sensitive files through the Bash<\/code> tool. You can write a simple hook to intercept Bash<\/code> commands that access blocked files, like this guard-bash-paths.py<\/code><\/a> hook.<\/p>\nPlugins<\/h2>\n
\n
~\/.claude\/settings.json<\/code> manually)<\/li>\n\/plugin-name:your-skill-name<\/code> prefix (no more conflicts)<\/li>\n<\/ul>\n\/plugin marketplace add openai\/codex-plugin-cc\n\/plugin marketplace add mattpocock\/skills\n\/plugin marketplace add vinta\/hal-9000\n\n# browse plugins\n\/plugin<\/code><\/pre>\n\n
Skills<\/h2>\n
# my skills\nnpx skills add https:\/\/github.com\/vinta\/hal-9000 --skill commit best-practices blindspot write-like-me refactor-agents-md -g\nnpx skills add https:\/\/github.com\/vinta\/dear-ai\n\n# workflow skills\nnpx skills add https:\/\/github.com\/mattpocock\/skills -g\n\n# writing skills\nnpx skills add https:\/\/github.com\/softaworks\/agent-toolkit --skill writing-clearly-and-concisely\nnpx skills add https:\/\/github.com\/hardikpandya\/stop-slop\nnpx skills add https:\/\/github.com\/shyuan\/writing-humanizer\n\n# doc skills\nnpx skills add https:\/\/github.com\/upstash\/context7 --skill find-docs -g\n\n# LLM API skills\nnpx skills add https:\/\/github.com\/openai\/skills\nnpx skills add https:\/\/github.com\/google-gemini\/gemini-skills\n\n# language skills\nnpx skills add https:\/\/github.com\/trailofbits\/skills --skill modern-python\nnpx skills add https:\/\/github.com\/trailofbits\/skills-curated --skill python-code-simplifier\nnpx skills add https:\/\/github.com\/dagster-io\/skills --skill dignified-python\nnpx skills add https:\/\/github.com\/wdm0006\/python-skills\n\n# backend skills\nnpx skills add https:\/\/github.com\/vintasoftware\/django-ai-plugins\nnpx skills add https:\/\/github.com\/google\/skills\nnpx skills add https:\/\/github.com\/cloudflare\/skills\nnpx skills add https:\/\/github.com\/supabase\/agent-skills\nnpx skills add https:\/\/github.com\/planetscale\/database-skills\n\n# frontend skills\nnpx skills add https:\/\/github.com\/vercel-labs\/agent-skills\nnpx skills add https:\/\/github.com\/vercel-labs\/next-skills\n\n# design skills\nnpx skills add https:\/\/github.com\/openai\/skills --skill frontend-skill\nnpx skills add https:\/\/github.com\/pbakaus\/impeccable\nnpx skills add https:\/\/github.com\/nextlevelbuilder\/ui-ux-pro-max-skill\nnpx skills add https:\/\/github.com\/Leonxlnx\/taste-skill\n\n# seo\/aeo skills\nnpx skills add https:\/\/github.com\/warpdotdev\/oz-skills --skill seo-aeo-audit\n\n# video skills\nnpx skills add https:\/\/github.com\/remotion-dev\/skills\n\n# browser skills\nnpx skills add https:\/\/github.com\/microsoft\/playwright-cli --skill playwright-cli -g\n\nnpx skills list -g\nnpx skills update -g\nnpx skills remove --all -g<\/code><\/pre>\n\n
\/brainstorming<\/code> from superpowers<\/a>: When in doubt, start with this skill<\/li>\n\/wayfinder<\/code> from mattpocock<\/a>: Let AI ask you a lot of questions<\/li>\n\/find-docs<\/code> from context7<\/a>: Find the latest documentations<\/li>\n\/frontend-design<\/code> from impeccable<\/a>: The better version of the official \/frontend-design<\/code> skill<\/li>\n\/simplify<\/code>: Run it often, you will like it<\/li>\n<\/ul>\nMCP Servers<\/h2>\n
Context7 MCP<\/h2>\n
ctx7<\/code> CLI with find-docs<\/code> skill instead.<\/p>\nnpx ctx7 setup<\/code><\/pre>\n\n
Playwright MCP<\/h3>\n
playwright-cli<\/code> skill instead. Both tools support headed mode<\/strong> (the opposite of headless), if you'd like to see the browser.<\/p>\nnpm install -g @playwright\/cli\nnpx skills add https:\/\/github.com\/microsoft\/playwright-cli<\/code><\/pre>\n\n
GitHub MCP<\/h3>\n
gh<\/code> command instead.<\/p>\nbrew install gh<\/code><\/pre>\nCodex MCP<\/h3>\n
codex exec<\/code> via CLI.<\/p>\n# Codex reads your local .codex\/config.toml by default\nclaude mcp add codex --scope user -- codex mcp-server<\/code><\/pre>\n\n
Hooks<\/h2>\n
SessionStart<\/code>, UserPromptSubmit<\/code>, and PreToolUse<\/code>.<\/p>\nPostToolUse<\/code> hook that runs deterministically:<\/p>\n{\n \"hooks\": {\n \"PostToolUse\": [\n {\n \"matcher\": \"Write|Edit\",\n \"hooks\": [\n {\n \"type\": \"command\",\n \"command\": \"uv run ruff check .\",\n \"if\": \"Edit(**\/*.py)\",\n \"timeout\": 30,\n \"statusMessage\": \"Linting Python code...\"\n }\n ]\n }\n ]\n }\n}<\/code><\/pre>\n\n
UserPromptSubmit<\/code> hook, which can set sessionTitle<\/code>.<\/p>\nUseful Tips<\/h2>\n
Prompt Best Practices<\/h3>\n
\n
Command Aliases<\/h3>\n
# in ~\/.zshrc\nalias cc=\"claude\"\nalias ccf=\"claude --model fable --effort max\"\nalias cct='tmux -CC new-session -s \"claude-$(date +%s)\" claude --teammate-mode tmux'\nalias ccy=\"claude --dangerously-skip-permissions\"\nccp() { claude --no-chrome --no-session-persistence -p \"$*\"; }<\/code><\/pre>\nccp<\/code> for ad-hoc prompts:<\/p>\nccp \"commit\"\nccp \"list all .md in this repo\"<\/code><\/pre>\n
![\"Claude](\"https:\/\/raw.githubusercontent.com\/vinta\/hal-9000\/main\/assets\/claude-code-statusline-grammar-check.png\")
<\/p>\n