Ponytail
Puts "the laziest senior developer at the company" inside Claude — forcing it to check if a feature already exists, is a browser default, or can be done in one line before writing anything.
Before & after
Same prompt, same agent — the only difference is the skill.
A date picker component is a good idea. You'd probably want something user-friendly with a calendar interface. There are libraries like React-DatePicker that work well. You could also consider Headless UI or Radix UI if you want more control over styling. You'd need to handle validation to make sure users enter valid dates. Accessibility is important too—keyboard navigation and screen reader support. Mobile users need touch-friendly interaction. You might need to add some styling to match your design. It's a fairly common requirement so there are lots of options. Would you like a basic version or something with more features?
SKIP THAT. The browser already has this.
<input type="date" name="birthdate" required />
Why:
→ Native date input handles calendar UI, mobile picker, locale formatting
→ Built-in validation (invalid dates rejected)
→ Keyboard accessible: tab + arrow keys work
→ Zero npm dependencies
→ 3 lines of code instead of 200 lines + bundle size
If you need custom styling for non-standard designs, add this instead of a library:
→ ::webkit-calendar-picker-indicator { color: var(--primary); }
Done. Ship it.About this skill
name: ponytail description: Use when Puts "the laziest senior developer at the company" inside Claude — forcing it to check if a feature already exists, is a browser default, or can be done in one line before writing anything.
Ponytail
Solves Claude's tendency to wildly over-engineer simple requests — asking for a date picker and getting a new library installed, a wrapper component built, and unsolicited timezone commentary. Before writing any code, Ponytail runs through a checklist: does this need to exist at all? Is it already in the codebase? Can it be done in one line? Only after passing all these checks does it implement the minimal version. In the date-picker example, without the skill Claude generated many lines across multiple approaches; with Ponytail on, it responded that the browser already has this built in and finished in one line. A /ponytail-review command can also be run against an existing project to trim excess.
What you get
- Pre-implementation minimalism checklist, dramatically shorter and more targeted code output, /ponytail-review to diet an existing codebase, meaningful token savings from not over-building.
Customize your output
- Adjust how strict the "does this need to exist" checklist is for your team's tolerance for abstraction.
Example output
Date picker request: without Ponytail, ~25 lines with a new library and wrapper component. With Ponytail on, 8 lines — reusing the browser's native date input.
Best for
Developers frustrated with Claude over-engineering simple requests who want genuinely minimal, necessary-only code.
SKILL.md preview
---
name: ponytail
description: Use this skill to force a minimal-implementation check before writing any code, confirming a feature does not already exist, is not a browser default, and cannot be done in one line.
version: 1.0.0
category: Development / Minimal Implementation
author: AgentVolt
license: proprietary
tags:
- development
- minimal-implementation
- standard
---
# Ponytail
Runs a strict minimal-implementation gate before writing any code: does this need to exist, is it already available, can it be done in one line.
## When to use
… (sign up to view the full skill)