Shorter technical answers, with code kept intact.
Shortens the agent's technical answers by removing filler while preserving technical substance and code.
When to use it
Use when the user requests caveman mode, fewer tokens, or brief answers, or invokes /caveman.
It changes how the agent writes responses, keeping the terse style active until the user explicitly turns it off.
Python enables the optional bundled compression, token-savings estimation, and response-linting tools; they use only the standard library.
No additional actions listed.