build-hook-librarylisted

# /build-hook-library — Hook Library > **Purpose.** Build a 50-100 hook library for a single subject (the agency operator OR a specific client). Each hook is tagged with: hook formula (one of 9 from `reference/frameworks/content/hook-anatomy.md`), pillar fit, voice-fit score, decay risk, last-used date. The library is the raw material every `/write-linkedin-post` and `/ghostwrite-client-post` pulls from. > > **Why this skill exists.** Without a pre-built, voice-calibrated, banned-vocab-clean library, posts default to whatever the runtime invents at write-time. Runtime defaults bias hard toward the patterns enumerated in `spec/BANNED-VOCABULARY.md` ("In this post, I'll share...", "Unpopular opinion:", emoji openers). The library shifts the default from generic-LinkedIn-mill to subject-voice-native. > > **Bar.** A linkedin-ghostwriter handed this library + a topic should be able to pick a hook in <60 seconds that scores ≥75 on `/score-hook` without rewrite. If they can't, the library is too thin or off-voice. ## When to invoke - After `/build-content-engine` completes for a new subject (agency-content or client-content) - Quarterly refresh trigger from `/audit-hook-library-decay` (when ≥40% of proven hooks have decayed below baseline) - New pillar added — pillar-fit gaps need filling - Voice profile re-extraction — hooks scored below 70 voice-fit need regeneration - New client onboarded — full library build for the client_slug ## Required inputs 1. Voice profile (≥70% — fr