The body is lean and assumes Claude's competence, never explaining what a benchmark or median is; every line earns its place, with only minor redundancy in the 'Boundaries' section restating the one-shot rule.

Provides a verbatim, copy-paste-ready ASCII scoreboard template with exact figures and an explicit rendering rule ('Render plain ASCII bars. The bar length shows the measured range; the label carries the exact figure'), which is concrete executable output for a display skill.

A single-purpose display skill with an unambiguous action ('Display this scoreboard when invoked') and a clear revert path ('stop ponytail'); per the simple-skill note, an unambiguous single action scores 3 without needing validation checkpoints.

Under 50 lines with no external bundle files needed and well-organized sections (Scoreboard, Honesty boundary, Boundaries), satisfying the simple-skill allowance for a top score on organization alone.

Enumerates concrete actions and measurable axes ('less code, less cost, more speed, from the benchmark medians') tied to a specific artifact (a compact scoreboard), matching the multiple-specific-actions anchor.

Explicitly answers both what (show measured impact as a scoreboard) and when (explicit 'Trigger:' clause), so it is not capped at 2 for a missing trigger clause.

Provides good coverage of natural user phrases plus the slash command ('/ponytail-gain', 'ponytail gain', 'what does ponytail save', 'show ponytail impact', 'ponytail scoreboard'), matching the good-coverage anchor.

The 'ponytail' namespace and distinct triggers give it a clear niche, and it explicitly distinguishes itself ('not a persistent mode, and not a per-repo number'), making conflict unlikely.