What it does
Dominik Tornow's hook scores the first 8,000 prompt characters, then adds a writing instruction. It cannot enforce a token limit; uncertain scores and errors leave the turn unchanged.
Agent tooling
stfu
Tell Claude Code when a short answer should do.
Bookmark: stfu
Keep this in your collection.
Leave a noteWhat would you try with this?
: stfu
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
How you can use it
If your coding assistant gives long answers to simple questions, study how this tool asks Jev to choose an answer length. It then adds an instruction to the assistant. That is advice, not a rule the assistant must obey. Uncertain scores leave the answer length alone.
It sends the first 8,000 characters of your request to TypeSafe's outside AI service. A developer would need to connect it to your assistant. The project has no licence, so get permission before copying or sharing its code.