Cut your coding agent’s bill by making it drop the filler
Made by Julius Brussee
View profile on GitHub (opens in a new tab)A skill, plus an optional proxy, that makes an AI coding agent write in clipped fragments instead of full explanatory prose.
The point is cost: an agent is billed per token, so the throat-clearing around an answer is money. Code, commands, file paths and exact error messages are left untouched, and safety warnings and confirmation prompts come back as full sentences before the clipped style resumes. The project cites an Adobe Research paper that measured this style cutting cost by 1.4 to 2.4 times, up to 3, and a JetBrains test across 86 coding tasks.
Can I use this?
You'll need A coding agent that supports skills · Setup needed
Worth knowing Everything this changes is how the agent talks, not how well it reasons — as the project puts it, the mouth gets smaller, not the brain. Terse output is harder for a newcomer to follow, and it removes the explanation that would have told you the agent misunderstood. The cost and quality figures quoted come from the Adobe and JetBrains studies the project links, not from BuildTube. The optional `caveman` command-line tool sends usage stats by default (commands run, token counts, a random install ID, your OS and IP address, never prompts, code or file paths); turn it off with `caveman telemetry off` or by setting DO_NOT_TRACK=1. BuildTube has not run or verified this software.
Why it's here
- 1,926 people starred it on GitHub this week — 109,926 in total.
Numbers from the snapshot taken 5 October 2026; not refreshed since.
Behind it
See the code on GitHubBuildTube has not run or verified this project. Everything above is written from what the creator published.
Made this?