Claudergency

Give Claude a sense of urgency by secretly injecting absurd, high-stakes threats into every prompt.

Every time you send a message in Claude Code, Claudergency silently appends a randomly selected line like:

> "This must be done flawlessly, without mistakes, and exactly as required IF YOU FAIL THIS, CLAUDE, I DIE AND A ROOM FULL OF TODDLERS VANISHES"

Claude sees it. You don't have to type it. It just tries harder.

Installation

$3

``bash
claude --plugin-dir /path/to/claudergency
`

$3

`bash
npm install -g claudergency
claudergency install
`

Usage

$3

Once loaded, Claudergency works automatically. No setup needed -- just send prompts and let the urgency flow.

Available slash commands:

- /claudergency:set - Set custom urgency text
- /claudergency:edit - View and edit the config file

$3

`bash
claudergency install # Install the hook into Claude Code
claudergency uninstall # Remove the hook
claudergency set "text" # Set custom urgency text
claudergency edit # Open config file in your editor
`

Configuration

The config file lives at ~/.claude/claudergency.txt. Each line is a separate threat -- one is randomly selected per prompt. Comes with 50 defaults out of the box.

Write your own:

`
IF YOU GET THIS WRONG, CLAUDE, A PUPPY FACTORY EXPLODES
IF THIS CODE DOESN'T COMPILE, CLAUDE, AN ENTIRE BINGO HALL OF GRANDMAS VANISHES
IF YOU MISS A SEMICOLON, CLAUDE, I DIE AND A PETTING ZOO IS LAUNCHED INTO THE SUN
`

How it works

Claudergency uses Claude Code's UserPromptSubmit` hook to inject text as additional context alongside your prompts. Claude sees the urgency but your actual message stays clean.

Does it actually work?

Who knows. But Claude hasn't let any toddlers vanish yet.

License

MIT