Lean body with no explaining of concepts Claude already knows; every line (commands, inline notes) earns its place, matching the efficient top anchor.

Provides fully executable cmux and git commands with clearly marked placeholders, copy-paste ready — meeting the executable-code top anchor rather than the pseudocode level below.

A clear 4-step sequence with an explicit verification checkpoint (read-screen expecting 'main -> main') before closing the pane, plus a conditional no-cmux branch; not the level below because validation is explicit.

Under 50 lines, single-purpose, no bundle files, with well-organized Steps and Notes sections — per the simple-skills scoring note this qualifies for the top anchor.

Lists multiple concrete actions — 'staging, committing, pushing, and cleaning up the cmux pane' — matching the top anchor for specific concrete actions.

Explicitly answers both what (commit/push skill changes to the private repo) and when ('Use after creating or updating any skill, when the user says...'), satisfying the explicit-trigger anchor.

Provides four natural phrases a user would actually say ('push the skill', 'push skills to github', 'save the skill to my repo', 'update the skills repo'), giving strong coverage; not the level below because it covers common variations rather than a single keyword.

Targets a clear niche — pushing agent skills to a specific private repo with distinct trigger phrases — making it unlikely to fire for unrelated skills.