ponytail-gain

Displays a one-shot ASCII scoreboard of benchmark-measured code, cost, and speed reductions.

Updated Sep 11, 2026
One-click install
npx skills add https://github.com/CYRUS-pinto/pi-bot --skill ponytail-gain-cyrus-pinto
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ponytail-gain
Source: https://github.com/CYRUS-pinto/pi-bot/tree/main/.agents/skills/ponytail-gain
Command: npx skills add https://github.com/CYRUS-pinto/pi-bot --skill ponytail-gain-cyrus-pinto

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Users want a quick, honest summary of ponytail's measured impact without digging through benchmark files or README documentation, and without misleading per-repo savings claims. ## Core Features & Use Cases - Benchmark Scoreboard: Renders a compact ASCII bar chart showing median reductions in lines of code (80-94%), cost (47-77%), and speed (3-6x faster) across 5 tasks and 3 models. - Honesty Boundary: Explicitly refuses to invent per-repo savings numbers, pointing instead to /ponytail-debt and /ponytail-audit for real per-repo figures. - One-Shot Display: Changes no mode, writes no files, and persists nothing; "stop ponytail" or "normal mode" reverts. - Use Case: A developer evaluating whether ponytail is worth adopting types "what does ponytail save" and immediately sees the published benchmark medians in a readable scoreboard. ## Quick Start Ask the assistant to show the ponytail gain scoreboard to display the benchmark median savings.

Frequently Asked Questions about ponytail-gain

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I see what ponytail saves in code and cost?

Invoke /ponytail-gain or ask "what does ponytail save" to render a one-shot ASCII scoreboard. It shows benchmark medians: 80-94% fewer lines of code, 47-77% lower cost, and 3-6x faster execution across 5 tasks and 3 models.

What benchmarks are the ponytail gain figures based on?

The figures are published benchmark medians from 5 everyday tasks (email validator, debounce, CSV sum, countdown timer, rate limiter) run across three models: Haiku, Sonnet, and Opus. They come from the benchmarks directory and README, not from your repository.

Can ponytail-gain show savings for my specific repository?

No, it deliberately refuses to print per-repo savings because the unbuilt version of your code was never written, so no real baseline exists. For actual per-repo figures, use /ponytail-debt for a counted ledger or /ponytail-audit for remaining cuttable code.

Does the ponytail gain scoreboard change my mode or write files?

No, it is strictly one-shot: it edits nothing, changes no mode, and writes no flag files. Saying "stop ponytail" or "normal mode" reverts any ponytail behavior separately.

When should I not rely on the ponytail gain numbers?

Do not treat the scoreboard as a prediction for your own repo, since the numbers are benchmark medians, not measurements of your codebase. Use it only for a general impact overview, and rely on /ponytail-debt for repo-specific accounting.