claude-brainrot — Claude Code skill: what it does, how to install
A skill that makes Claude drop memes into every reply: image-plus-sound combos and standalone sounds. It loads itself, no manual call.
npx skills add marks97/claude-brainrot --skill claude-brainrot -a claude-codeWhat claude-brainrot does
The skill hooks into UserPromptSubmit, so it fires on every message you send. The hook nudges the agent, and the agent picks a meme that fits the moment: a short answer gets one image with sound or a couple of sound-only hits, a long multi-step response gets up to three images and five sound fires. There is no skip turn.
Inside the skill folder sits catalogue.json with two pools, images and sounds. Every entry carries a slug, a path, tags and a vibe field describing when to use it. The tags are plain: bruh, rage, laugh, alert, shock, confusion. The agent reads the image itself, while the hook picks the paired sound using the lock_sound, silent and preferred_sounds flags. Sounds marked standalone only play on their own.
Playback goes through afplay, so it will not work on Linux or Windows — worth checking before you install. Two more hooks sit in the skill: PreToolUse blocks reading images from the command line, and PostToolUse starts the paired sound. For an image to render inline in chat, its Read call needs a short line of text before and after, otherwise the UI collapses it into a dropdown.
What it helps with
- break up a dry answer with a meme
- match an image to the mood of the chat
- drop a sound between tool calls
- avoid repeating the same meme in one session
- play audio in the background without blocking work
Example prompts for the agent
- Answer the question and drop a meme that fits the mood
- Add a sound when the build finishes
- Something broke, fire a meme tagged rage
How to install claude-brainrot
1Skills CLI (Vercel)
npx skills add marks97/claude-brainrot --skill claude-brainrot -a claude-code2Manually: put the files in place
~/.claude/skills/claude-brainrot/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/claude-brainrot/
1Skills CLI (Vercel)
npx skills add marks97/claude-brainrot --skill claude-brainrot -a codex2Manually: put the files in place
.agents/skills/claude-brainrot/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add marks97/claude-brainrot --skill claude-brainrot -a cursor2Manually: put the files in place
.cursor/skills/claude-brainrot/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install marks97/claude-brainrot claude-brainrot2Manually: put the files in place
.github/skills/claude-brainrot/SKILL.md1Skills CLI (Vercel)
npx skills add marks97/claude-brainrot --skill claude-brainrot -a gemini-cli2Platform CLI
gemini skills install https://github.com/marks97/claude-brainrotManually: put the files in place
.devin/skills/claude-brainrot/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add marks97/claude-brainrot --skill claude-brainrot -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about claude-brainrot
How do I install claude-brainrot in Claude?
Run npx skills add marks97/claude-brainrot --skill claude-brainrot -a claude-code. The skill lands at ~/.claude/skills/claude-brainrot/SKILL.md.
What does claude-brainrot do?
On every message, the UserPromptSubmit hook tells the agent to drop a meme: an image with a paired sound or a sound-only fire. How many depends on the length of the response.
Do I need to invoke the skill manually?
No. It loads and stays active on its own, with no slash command or arguments. The agent decides when a meme fits.
