Ponytail makes your coding agent write less code

Most AI coding agents write too much code. Ponytail is a skill that makes them write less.

It's one prompt that works with Claude Code, Codex and a long list of other agents (the repo says 20). Before writing anything, the agent has to climb a short ladder. Does this need to exist? Is it already in the codebase? Does the standard library or the browser already do it? Only then does it write the minimum that works, plus one small test if there's real logic.

The README's example is a date picker. Without the skill, the agent built one by hand in 335 lines. With it, the agent reused the repo's existing input component and the browser's native date picker. Ten lines.

The author's own benchmark (39 tasks, Opus 5.5, 5 runs each) reports about 53% less code, and 98% of risky logic shipping with a test versus 68% without. Those are the author's numbers, so try it on your own repo before you trust them.

I run it on every coding task. It also has /ponytail-review and /ponytail-audit, which hunt for code you can delete.

Which skill do you install first on a new machine?